New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

linuxkit bake #2648

Open
wants to merge 9 commits into
base: master
from

Conversation

Projects
None yet
6 participants
@w9n
Copy link
Contributor

w9n commented Oct 28, 2017

When writing a moby yml I do not care about the container version and hashes (only if I specifically want to stay with a version) and would like to simplify the process of writing them. I also do not want to keep track of all pkg version I use downstream and update them manually when something changes upstream.

The idea is to optionally split the moby config into 1 mutable and n immutable versions. The mutable version describes the desired state which uses image names like linuxkit/containerd:mutable(or the usual hash tags for pinning down on a version), while the immutable configs stay like currently and get created on demand or mutation from a specifc pkg git repo. They can be versioned and stored in the end.

Therefore linuxkit pkg collect --pkgroot /path/to/repo/linuxkit mutable.yml first collects which container in the mutable yml are tagged mutable and then walks recursivly through pkgroot and looks for build.ymlof a package with the same image name. When found it calculates and stores the package hash (Not sure about error case). At this Point we could compare with the last hash of the image and implement build and push on change. At the end a yml with the hashes gets written versioned into a specific dir near the mutable yml and persist as an Artifact for a possible build.

This could also be used for CI that triggers linuxkit pkg collect --push on every pull for every mutable image yml. Would need to save the immutable artifacts somewhere.

The big disadvantage is that you probably dont want moby as dependency in linuxkit which is needed to parse the moby config (On the other side the linuxkit git pkg for the hashes is private). Let me know what you think about it!

(Im just getting used to go, so its a bit messy and uncomplete yet)

@GordonTheTurtle

This comment has been minimized.

Copy link
Collaborator

GordonTheTurtle commented Oct 28, 2017

Please sign your commits following these rules:
https://github.com/moby/moby/blob/master/CONTRIBUTING.md#sign-your-work
The easiest way to do this is to amend the last commit:

$ git clone -b "pkg_collect" git@github.com:w9n/linuxkit.git somewhere
$ cd somewhere
$ git commit --amend -s --no-edit
$ git push -f

Amending updates the existing PR. You DO NOT need to open a new one.

@rn

This comment has been minimized.

Copy link
Member

rn commented Oct 28, 2017

I don't understand the explanation for this PR (maybe I'm missing some context).

Could you first describe which problem you try to solve?

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Oct 28, 2017

added some context to the description.

@deitch

This comment has been minimized.

Copy link
Collaborator

deitch commented Oct 29, 2017

When writing a moby yml I do not care about the container version and hashes (only if I specifically want to stay with a version) and would like to simplify the process of writing them. I also do not want to keep track of all pkg version I use downstream and update them manually when something changes upstream

I have a hard time understanding how that would work with immutable images (which is what LinuxKit is all about).

I think this might be easier to understand if you gave sample .yml files for what it would look like if this were in place?

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Oct 29, 2017

i basically thought of a not buildable config

kernel:
  image: linuxkit/kernel:4.9.58
  cmdline: "console=tty0 console=ttyS0 console=ttyAMA0"
init:
  - linuxkit/init:mutable
  - linuxkit/runc:mutable
  - linuxkit/containerd:82be2bbb7cf83bab161ffe2a64624ba1107725ff
  - linuxkit/ca-certificates:mutable
onboot:
  - name: sysctl
    image: linuxkit/sysctl:mutable
  - name: rngd1
    image: linuxkit/rngd:mutable
    command: ["/sbin/rngd", "-1"]
services:
  - name: getty
    image: linuxkit/getty:mutable
    env:
     - INSECURE=true
  - name: rngd
    image: linuxkit/rngd:mutable
  - name: dhcpcd
    image: linuxkit/dhcpcd:mutable
  - name: sshd
    image: linuxkit/sshd:mutable
files:
  - path: root/.ssh/authorized_keys
    source: ~/.ssh/id_rsa.pub
    mode: "0600"
    optional: true
trust:
  org:
    - linuxkit

when used with linuxkit pkg collect --pkgroot /path/to/repo/linuxkit ssh-mutable.yml creates a new file artifacts/ssh-vXYZ.yml with the most current hashes of pkgroot

kernel:
  image: linuxkit/kernel:4.9.58
  cmdline: "console=tty0 console=ttyS0 console=ttyAMA0"
init:
  - linuxkit/init:42a92119e1ca10380e0d33e26c0cbcf85b9b3558
  - linuxkit/runc:817fdc592eac6cb7804fa1721a43a7f6e23fb50f
  - linuxkit/containerd:82be2bbb7cf83bab161ffe2a64624ba1107725ff
  - linuxkit/ca-certificates:af4880e78edc28743f7c5e262678c67c6add4c26
onboot:
  - name: sysctl
    image: linuxkit/sysctl:a9ad57ed738a31ea9380cd73236866c312b35489
  - name: rngd1
    image: linuxkit/rngd:842e5e8ece7934f0cab9fd0027b595ff3471e5b9
    command: ["/sbin/rngd", "-1"]
services:
  - name: getty
    image: linuxkit/getty:626ccc8e1766c40447f29a790d3a7cfff126f2a2
    env:
     - INSECURE=true
  - name: rngd
    image: linuxkit/rngd:842e5e8ece7934f0cab9fd0027b595ff3471e5b9
  - name: dhcpcd
    image: linuxkit/dhcpcd:48831507404049660b960e4055f544917d90378e
  - name: sshd
    image: linuxkit/sshd:f55ec010619e19178d5daecb4e595e84ecbf7d67
files:
  - path: root/.ssh/authorized_keys
    source: ~/.ssh/id_rsa.pub
    mode: "0600"
    optional: true
trust:
  org:
    - linuxkit

I think the main practical reason is to use it in CI where it is necessary to know which pkg has changed to not rebuild everything on a merge. It also gives gives rollback capabilities outside the git history.

But I reevaluated priorities and think its a nice to have but linuxkit is working fine now and i dont think its worth the moby dep. I might create a little external tool for practicing, but low priority.

There is a different problem that a container linuxkit/sshd:f55ec010619e19178d5daecb4e595e84ecbf7d67 is not really immutable because it pulls latest alpine packages on container build time. The current Hash is needed for multi arch but by creating ymls as artifacts it might be possible to use real hashes of a container and make the yml artifacts platform specific. But this gets out of scope for me.

@deitch

This comment has been minimized.

Copy link
Collaborator

deitch commented Oct 29, 2017

So, is this basically a way of saying, "for some packages, I want the latest available hash at build time"?

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Oct 29, 2017

yep, but still create the immutable ymls so its reproducable what has been build and compare what has been changed.

@deitch

This comment has been minimized.

Copy link
Collaborator

deitch commented Oct 29, 2017

So rather than the usual moby .yml, it kind of is a moby "template" .yml, that gets parsed to generate the real .yml with immutable hashes? Still immutable, but somewhat easier to maintain?

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Oct 29, 2017

exactly. It would enable other possibilities as well as described above but there might be other ideas for CI and real hashes.

@deitch

This comment has been minimized.

Copy link
Collaborator

deitch commented Oct 29, 2017

OK, I get it now; want to think it over somewhat. @rn had commented earlier; let's see his thoughts.

I am going to rename the issue.

@deitch deitch changed the title WIP: linuxkit pkg collect WIP: linuxkit pkg template yml to generate immutable yml for reuse Oct 29, 2017

@rn

This comment has been minimized.

Copy link
Member

rn commented Oct 29, 2017

This would be better done in the moby tool as it handles the build YAML files. There are two related proposals:

@deitch

This comment has been minimized.

Copy link
Collaborator

deitch commented Oct 29, 2017

Ah yes, the .mobytags and overrides proposals.

@w9n does that address the question?

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Oct 29, 2017

yes, it pretty much goes is in the right direction. I will read and think about the related Issues and PRs in the next days.

@w9n w9n closed this Oct 29, 2017

@ijc

This comment has been minimized.

Copy link
Collaborator

ijc commented Nov 1, 2017

I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this one yet, but one quick though I had was that what is called foo:mutable here would be called foo:latest in many other contexts.

@w9n w9n referenced this pull request Nov 2, 2017

Closed

Proposal: package managment #2674

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Nov 2, 2017

i didnt choose foo:latest because it can conflict with other contexts. But can maybe be prevented if carefully designed. I created a proper proposal.

@w9n w9n reopened this Dec 17, 2017

@w9n w9n force-pushed the w9n:pkg_collect branch from a07bb35 to 9ba959b Dec 17, 2017

@GordonTheTurtle GordonTheTurtle removed the dco/no label Dec 17, 2017

@w9n w9n force-pushed the w9n:pkg_collect branch 3 times, most recently from 3ebd47f to 557bb5c Dec 17, 2017

@w9n w9n referenced this pull request Dec 17, 2017

Merged

add omitempty for yml config #194

@w9n

This comment has been minimized.

Copy link
Contributor

w9n commented Dec 17, 2017

Since moby became a dependency this works now depending on moby/tool#194 linuxkit pkg collect --pkgroot ~/go/src/github.com/linuxkit/linuxkit/pkg test.yml

test,yml
kernel:
  image: linuxkit/kernel:4.9.58
  cmdline: "console=tty0 console=ttyS0 console=ttyAMA0"
init:
  - linuxkit/init:<latest>
  - linuxkit/runc:<latest>
  - linuxkit/containerd:82be2bbb7cf83bab161ffe2a64624ba1107725ff
  - linuxkit/ca-certificates:<latest>
onboot:
  - name: sysctl
    image: linuxkit/sysctl:<latest>
  - name: rngd1
    image: linuxkit/rngd:<latest>
    command: ["/sbin/rngd", "-1"]
services:
  - name: getty
    image: linuxkit/getty:<latest>
    env:
     - INSECURE=true
  - name: rngd
    image: linuxkit/rngd:<latest>
  - name: dhcpcd
    image: linuxkit/dhcpcd:<latest>
  - name: sshd
    image: linuxkit/sshd:<latest>
files:
  - path: root/.ssh/authorized_keys
    source: ~/.ssh/id_rsa.pub
    mode: "0600"
    optional: true
trust:
  org:
    - linuxkit
result hash of the yml files bytes (5f39869eef39b525e90661e2327ed645c1a9df8a.yml)
kernel:
  image: linuxkit/kernel:4.9.58
  cmdline: console=tty0 console=ttyS0 console=ttyAMA0
init:
- linuxkit/init:9250948d0de494df8a811edb3242b4584057cfe4
- linuxkit/runc:abc3f292653e64a2fd488e9675ace19a55ec7023
- linuxkit/containerd:82be2bbb7cf83bab161ffe2a64624ba1107725ff
- linuxkit/ca-certificates:de21b84d9b055ad9dcecc57965b654a7a24ef8e0
onboot:
- name: sysctl
  image: linuxkit/sysctl:ce3bde5118a41092f1b7048c85d14fb35237ed45
- name: rngd1
  image: linuxkit/rngd:94e01a4b16fadb053455cdc2269c4eb0b39199cd
  command:
  - /sbin/rngd
  - "-1"
onshutdown: []
services:
- name: getty
  image: linuxkit/getty:22e27189b6b354e1d5d38fc0536a5af3f2adb79f
  env:
  - INSECURE=true
- name: rngd
  image: linuxkit/rngd:94e01a4b16fadb053455cdc2269c4eb0b39199cd
- name: dhcpcd
  image: linuxkit/dhcpcd:0d59a6cc03412289ef4313f2491ec666c1715cc9
- name: sshd
  image: linuxkit/sshd:ac5e8364e2e9aa8717a3295c51eb60b8c57373d5
trust:
  org:
  - linuxkit
files:
- path: root/.ssh/authorized_keys
  directory: false
  source: ~/.ssh/id_rsa.pub
  optional: true
  mode: "0600"

_
I would like to have a file like ~/.linuxkit.conf where i can specify multiple pkg roots.
It could also easily be upgraded to support version tags in that file (as proposed in moby/tool#101 ) or in the build.yml in the pkg directory.

@w9n w9n force-pushed the w9n:pkg_collect branch from 557bb5c to 083d497 Dec 26, 2017

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 18, 2019

datakit
20 items modified
Updated
{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358:dco-signed[success] descr=All commits are signed}
{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1322?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 3272[c28d72169dc11c444be719b5bf3c7134ace09358] master rn open "Update kernels to 4.20.3/4.19.16/4.14.94/4.9.151" 
[|0 rn: "![commander butterfly caterpillar](https://user-images.githubusercontent.com/333"|]}
{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358} 
Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 18, 2019

datakit
21 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed
{{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358} ci/circleci}
{{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358} ci/datakit/linuxkit-ci}
{{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358} dco-signed}
{linuxkit/linuxkit 3272}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit c28d72169dc11c444be719b5bf3c7134ace09358}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 18, 2019

datakit
21 items modified
Updated
{linuxkit/linuxkit heads/master[5cf518b89bbe5fa063591e05d0c374f58d1d7e39]}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 5cf518b89bbe5fa063591e05d0c374f58d1d7e39} 
Removed
{{linuxkit/linuxkit eeb2d546d1b8920161894aafc4b9ef842b80e29d} ci/circleci}
{{linuxkit/linuxkit eeb2d546d1b8920161894aafc4b9ef842b80e29d} ci/datakit/linuxkit-ci}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eeb2d546d1b8920161894aafc4b9ef842b80e29d}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 18, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 5cf518b89bbe5fa063591e05d0c374f58d1d7e39:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1323?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests are queued behind your running builds}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 18, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 5cf518b89bbe5fa063591e05d0c374f58d1d7e39:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1323?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
22 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3254[30b8670a98bab7a5bd3d343ad2613e3ed783e8e5] master errordeveloper open "Add runtime.envFromFiles" 
[|0 errordeveloper: "";
  450501032 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  450522326 errordeveloper: "Just a few thoughts on my contributor experience, hopefully useful :)";
  450809805 errordeveloper: "@justincormack I've updated this. Let me know if you want me to open another PR "|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 30b8670a98bab7a5bd3d343ad2613e3ed783e8e5} 
Removed
{{linuxkit/linuxkit b7268fae5a2bac56d82a01dce71977a6cdcbb790} ci/circleci}
{{linuxkit/linuxkit b7268fae5a2bac56d82a01dce71977a6cdcbb790} ci/datakit/linuxkit-ci}
{{linuxkit/linuxkit b7268fae5a2bac56d82a01dce71977a6cdcbb790} dco-signed}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit b7268fae5a2bac56d82a01dce71977a6cdcbb790}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 30b8670a98bab7a5bd3d343ad2613e3ed783e8e5:dco-signed[success] descr=All commits are signed}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 30b8670a98bab7a5bd3d343ad2613e3ed783e8e5:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1324?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests are queued behind your running builds}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 30b8670a98bab7a5bd3d343ad2613e3ed783e8e5:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1324?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
19 items modified
Updated
{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1326?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests are queued behind your running builds}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a} 
Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a:dco-signed[success] descr=All commits are signed}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1326?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
18 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3254[30b8670a98bab7a5bd3d343ad2613e3ed783e8e5] master errordeveloper open "Add runtime.envFromFiles" 
[|0 errordeveloper: "";
  450501032 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  450522326 errordeveloper: "Just a few thoughts on my contributor experience, hopefully useful :)";
  450809805 errordeveloper: "@justincormack I've updated this. Let me know if you want me to open another PR ";
  455766390 errordeveloper: "I've rebased this."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and "|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just "|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just ";
  455805148 errordeveloper: "I've not been able to boot the 4.19 one (getting meaningless error from hyperkit"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
19 items modified
Updated
{linuxkit/linuxkit d6a94f30cdbd2b94bcebcc07c31de516432d2751:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1327?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests are queued behind your running builds}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 3274[d6a94f30cdbd2b94bcebcc07c31de516432d2751] master yankcrime open "[OpenStack] Support os-client-config usage for authentication" 
[|0 yankcrime: "**- What I did**\r\n\r\nUpdated the [Gophercloud](https://github.com/gophercloud/gop"|]}
{linuxkit/linuxkit d6a94f30cdbd2b94bcebcc07c31de516432d2751} 
Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit d6a94f30cdbd2b94bcebcc07c31de516432d2751:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1327?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=CircleCI is running your tests}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
22 items modified
Updated
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1328?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests are queued behind your running builds}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 3274[e91168b65d4278f23dbb2916f72d3abdd24aa82b] master yankcrime open "[OpenStack] Support os-client-config usage for authentication" 
[|0 yankcrime: "**- What I did**\r\n\r\nUpdated the [Gophercloud](https://github.com/gophercloud/gop"|]}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b} 
Removed
{{linuxkit/linuxkit d6a94f30cdbd2b94bcebcc07c31de516432d2751} ci/circleci}
{{linuxkit/linuxkit d6a94f30cdbd2b94bcebcc07c31de516432d2751} ci/datakit/linuxkit-ci}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit d6a94f30cdbd2b94bcebcc07c31de516432d2751}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b:ci/circleci[pending] url=https://circleci.com/gh/linuxkit/linuxkit/1328?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=CircleCI is running your tests}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
17 items modified
Updated
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1328?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 19, 2019

datakit
23 items modified
Updated
{linuxkit/linuxkit 1df6804d99a39599c82c8aedc584d05374204a43:dco-signed[success] descr=All commits are signed}
{linuxkit/linuxkit 1df6804d99a39599c82c8aedc584d05374204a43:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1329?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3274[1df6804d99a39599c82c8aedc584d05374204a43] master yankcrime open "[OpenStack] Support os-client-config usage for authentication" 
[|0 yankcrime: "**- What I did**\r\n\r\nUpdated the [Gophercloud](https://github.com/gophercloud/gop"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 1df6804d99a39599c82c8aedc584d05374204a43} 
Removed
{{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b} ci/circleci}
{{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b} ci/datakit/linuxkit-ci}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 20, 2019

datakit
18 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just ";
  455805148 errordeveloper: "I've not been able to boot the 4.19 one (getting meaningless error from hyperkit";
  455872321 rn: "Last time I looked at ZFS one issue was that the version of the tools which came"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 20, 2019

datakit
18 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just ";
  455805148 errordeveloper: "I've not been able to boot the 4.19 one (getting meaningless error from hyperkit";
  455872321 rn: "Last time I looked at ZFS one issue was that the version of the tools which came";
  455895367 errordeveloper: "The userspace compatibility is something we should confirm, I shall play\naround "|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 21, 2019

datakit
18 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just ";
  455805148 errordeveloper: "I've not been able to boot the 4.19 one (getting meaningless error from hyperkit";
  455872321 rn: "Last time I looked at ZFS one issue was that the version of the tools which came";
  455895367 errordeveloper: "The userspace compatibility is something we should confirm, I shall play\naround ";
  455981118 rn: "> (getting meaningless error from hyperkit)\r\n\r\nWhat type of errors? Might be rel"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 21, 2019

datakit
18 items modified
Updated
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just ";
  455805148 errordeveloper: "I've not been able to boot the 4.19 one (getting meaningless error from hyperkit";
  455872321 rn: "Last time I looked at ZFS one issue was that the version of the tools which came";
  455895367 errordeveloper: "The userspace compatibility is something we should confirm, I shall play\naround ";
  455981118 rn: "> (getting meaningless error from hyperkit)\r\n\r\nWhat type of errors? Might be rel";
  456018564 errordeveloper: "@rn yes, I think I got that same error, but with the staircase effect :)\r\n\r\nShou"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed  {linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 21, 2019

datakit
23 items modified
Updated
{linuxkit/linuxkit heads/master[9a818a1446d67875ceb30aa02b2b555ce05a970c]}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3273[872a92849b64b2137a93249f8884e225ea9b549a] master errordeveloper open "Use latest stable release of ZFS" 
[|0 errordeveloper: "This change makes `make -C kernel build_zfs_4.14.x` succeed.\r\n\r\nCloses #2950.";
  455766576 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  455768281 justincormack: "Does this work on 4.19? Given we switched default kernel to 4.19 maybe makes sen";
  455770310 errordeveloper: "I've tested `make -C kernel build_zfs_4.9.x` just now, and that worked. Will try";
  455770690 errordeveloper: "By the way, they have [`0.8.0-rc3`](https://github.com/zfsonlinux/zfs/tree/zfs-0";
  455770842 errordeveloper: "@justincormack whatever is the outcome of this, should we enable ZFS in CI, and ";
  455776688 errordeveloper: "@justincormack `make ORG=errordeveloper NOTRUST=1 build_zfs_4.19.x` worked just ";
  455805148 errordeveloper: "I've not been able to boot the 4.19 one (getting meaningless error from hyperkit";
  455872321 rn: "Last time I looked at ZFS one issue was that the version of the tools which came";
  455895367 errordeveloper: "The userspace compatibility is something we should confirm, I shall play\naround ";
  455981118 rn: "> (getting meaningless error from hyperkit)\r\n\r\nWhat type of errors? Might be rel";
  456018564 errordeveloper: "@rn yes, I think I got that same error, but with the staircase effect :)\r\n\r\nShou"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}
{linuxkit/linuxkit 9a818a1446d67875ceb30aa02b2b555ce05a970c} 
Removed
{{linuxkit/linuxkit 5cf518b89bbe5fa063591e05d0c374f58d1d7e39} ci/circleci}
{{linuxkit/linuxkit 5cf518b89bbe5fa063591e05d0c374f58d1d7e39} ci/datakit/linuxkit-ci}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 5cf518b89bbe5fa063591e05d0c374f58d1d7e39}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 21, 2019

datakit
23 items modified
Updated
{linuxkit/linuxkit 9a818a1446d67875ceb30aa02b2b555ce05a970c:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1330?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment linuxkit/linuxkit#2648 (comment)";
  356587899 w9n: "> Do you have an example for a baked YAML file. It would be useful if it would c";
  356589814 w9n: "i dont really trust the test time, something must be wrong...\r\n```\r\n[+ 24m 58s] ";
  356647166 w9n: "@ijc \r\n```\r\nlinuxkit bake -D foo=abcdef foo.yml\r\n```\r\nThis can actually be done ";
  356922250 w9n: "updated the last comment, otherwise i guess i should revert setting global confi";
  357164648 w9n: "Sorry for all these changes\r\n - dropped putting GlobalConfig in its own namespac";
  357425907 w9n: "```\r\n[+ 24m 27s] [PASS    ] linuxkit.build.examples.bake 32.57s\r\n```\r\n\r\nfixed te"|]}

Removed
{{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a} ci/circleci}
{{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a} ci/datakit/linuxkit-ci}
{{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a} dco-signed}
{linuxkit/linuxkit 3273}
{linuxkit/linuxkit 0748457296dda072557f13358346261a8c1948b8}
{linuxkit/linuxkit 13bfb5c238b453bdf578b7c11c3055ef4074505d}
{linuxkit/linuxkit 1616b18fb07ba39d04ddd616593769b2d0384fef}
{linuxkit/linuxkit 1721ea26718943ca036f036fad894b0f05032e77}
{linuxkit/linuxkit 38b563a11443db7e81211a9e7fd8717c941178d0}
{linuxkit/linuxkit 4b66ef2e2a65a6c8bb2c03fd95fc09527b3f5c5f}
{linuxkit/linuxkit 4e634e3b2975c36450be70d550080b76f71a9c06}
{linuxkit/linuxkit 618c8675e364d936ee7561f411365e2df217806c}
{linuxkit/linuxkit 872a92849b64b2137a93249f8884e225ea9b549a}
{linuxkit/linuxkit 89a7fbebcc83c2b1b2a4ffce29a8f65d11fd540b}
{linuxkit/linuxkit a36e1149a9b6daffeb8778e87b0b5f490752858a}
{linuxkit/linuxkit bbde423c6f9190c7b32ff2974422de0daa59cbd1}
{linuxkit/linuxkit d45f905adb8623c6c7ca06b3b4c5e3c66f4882e7}
{linuxkit/linuxkit e91168b65d4278f23dbb2916f72d3abdd24aa82b}
{linuxkit/linuxkit eef5d1a9051a7db6d08e70d4d3c8202d977b36ff}
{linuxkit/linuxkit ef9302bc01f45e937eef52f0a68d3b9f073e0747}

datakit pushed a commit to linuxkit/linuxkit-logs that referenced this pull request Jan 23, 2019

datakit
21 items modified
Updated
{linuxkit/linuxkit a94099639ec2956ec30ccc566128887131b7ff4f:dco-signed[success] descr=All commits are signed}
{linuxkit/linuxkit a94099639ec2956ec30ccc566128887131b7ff4f:ci/circleci[success] url=https://circleci.com/gh/linuxkit/linuxkit/1331?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link descr=Your tests passed on CircleCI!}
{linuxkit/linuxkit 3168[1d1c8fcee99d58781b90bc646398046d22e286ef] master kkroo open "Use containerd restart manager to monitor services" 
[|0 kkroo: "<!--\r\nPlease make sure you've read and understood our contributing guidelines;\r\n";
  416053776 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  416566165 justincormack: "Hi, we can't merge PRs unless they are signed-off-by as per instructions above.\r";
  417360415 ijc: "containerd/containerd#2318 shows it using `container.Upd";
  442205572 kkroo: "Addressing comments regarding race conditions. Regarding the use of this interfa";
  442222157 medic15: "This needs to be configurable for each service.  Not all those who wander are lo";
  442319928 kkroo: "> This needs to be configurable for each service. Not all those who wander are l";
  442400886 ijc: "@medic15 and @justincormack are correct, this needs to be configurable per servi";
  442401015 justincormack: "At present a few of the services we run may terminate (eg `rngd` if no rng is su";
  442472675 medic15: "My use case involves populating a Redis in-memory database using data from sever";
  442473416 justincormack: "@medic15 yes that seems a reasonable use case where it should be configurable.";
  442621437 kkroo: "@medic15 @justincormack I still don't understand the use case and why such a one";
  442622363 justincormack: "Services are not started until all `onboot` containers have exited."|]}
{linuxkit/linuxkit 3276[a94099639ec2956ec30ccc566128887131b7ff4f] master zx2c4 open "wireguard: upgrade to 0.0.20190123" 
[|0 zx2c4: "* tools: curve25519: handle unaligned loads/stores safely\r\n\r\nThis should fix spo"|]}
{linuxkit/linuxkit 2648[bfaa9e7fc599a8b9edba4429ae513f0736411fce] master w9n open "linuxkit bake" 
[|0 w9n: "When writing a moby yml I do not care about the container version and hashes (on";
  340192010 GordonTheTurtle: "<!-- AUTOMATED:POULE:DCO-EXPLANATION -->\nPlease sign your commits following thes";
  340204660 rn: "I don't understand the explanation for this PR (maybe I'm missing some context).";
  340224169 w9n: "added some context to the description. ";
  340241991 deitch: "> When writing a moby yml I do not care about the container version and hashes (";
  340268573 w9n: "i basically thought of a not buildable config\r\n```\r\nkernel:\r\n  image: linuxkit/k";
  340269095 deitch: "So, is this basically a way of saying, \"for some packages, I want the latest ava";
  340269511 w9n: "yep, but still create the immutable ymls so its reproducable what has been build";
  340269588 deitch: "So rather than the usual moby `.yml`, it kind of is a moby \"template\" `.yml`, th";
  340269816 w9n: "exactly. It would enable other possibilities as well as described above but ther";
  340269924 deitch: "OK, I get it now; want to think it over somewhat. @rn had commented earlier; let";
  340270518 rn: "This would be better done in the `moby` tool as it handles the build YAML files.";
  340271020 deitch: "Ah yes, the `.mobytags` and overrides proposals.\r\n\r\n@w9n does that address the q";
  340273985 w9n: "yes, it pretty much goes is in the right direction. I will read and think about ";
  341074173 ijc: "I'm still catching up on my post PTO backlog(s) so I haven't fully grokked this ";
  341552432 w9n: "i didnt choose ```foo:latest``` because it can conflict with other contexts. But";
  352289887 w9n: "Since moby became a dependency this works now depending on moby/tool#194 `linuxk";
  355176335 w9n: "@deitch @rn @ijc Please recheck, adopting this could quit manually updating hash";
  355233916 deitch: "I remain torn on this. I think I still like the idea of having the ability to ta";
  355287913 w9n: "> If I understand the implementation correctly, pkg collect would then look in l";
  355293548 ijc: "> if a container has `<latest>` tag\r\n\r\nAre the angle-brackets (`<>`) literally p";
  355298378 w9n: "> Are the angle-brackets (<>) literally part of the tag here or just \"syntax\"?\r\n";
  355304180 deitch: "> Yes, you would need atleast the linuxkit/linuxkit repo and a setup pkgroot par";
  355307237 w9n: "> Are there 2 distinct behaviours?\r\n\r\nThere is `linuxkit pkg collect --pkgroot` ";
  355310745 deitch: "> there is linuxkit build -collect which pre-processes and directly outputs the ";
  355311332 rn: "`linuxkit pkg collect` seems a misnomer as it is not an operation on a package. ";
  355313071 deitch: "> linuxkit pkg collect seems a misnomer as it is not an operation on a package\r\n";
  355314861 rn: "> Yes, but I could see how you might want a convenience of a single step.\r\n\r\nYou";
  355316644 deitch: ">  I really do not want linuxkit build to do too much magic. It should just buil";
  355316963 deitch: "On second thought, I guess I can see it. `lkt` already contains all of the code ";
  355318030 rn: "sure, this can be done with a shell script. `grep` image patterns and `linuxkit ";
  355321845 ijc: "> You get a \"single\" step with linuxkit tranform | linuxkit build -. I really do";
  355347992 w9n: "dropping `linuxkit build -collect` and `--pkgroot <org>:<path>` sounds good to m";
  355446448 rn: "`linuxkit transform -o outputfile <filename> | linuxkit build -` does not seem t";
  355491579 deitch: ">  I think generate is my favourite\r\n\r\nI like it too. My only concern is that, t";
  355491607 deitch: "Or perhaps `lkt template`?";
  355492212 w9n: "generate sounds good but dont mind changing.\r\n\r\nChanged to `linuxkit generate --";
  355530363 rn: "@w9n thanks for the quick change, I have a few comments on the implementation it";
  355535990 ijc: "> then a new `linuxkit build` command \r\n\r\nITYM `linuxkit generate`? Not mad enam";
  355538317 ijc: "> My only concern is whether or not there is some existing YAML templating synta";
  355540809 deitch: "> Couldn't find anything\r\n  \r\nYeah, I had that issue a while ago. Ended up writi";
  355540982 deitch: "> ITYM linuxkit generate? Not mad enamoured on the name, but no better ideas (ba";
  355567882 w9n: "> you shouldn't need to add anything to pkglib. This is a new top level command ";
  355569715 w9n: "bake & build would kind of short and precisely describe whats going on and actua";
  355570963 ijc: "> the git methods are not public \r\n\r\nYou appear to be using them to reimplement ";
  355572091 w9n: "Yes `New` in pkglib.Pkg would be awesome - but i think this needs some more work";
  355574550 ijc: "Where/why does this code need to deal with old hashes?\r\n\r\nGenerating the hash co";
  355577199 w9n: "if there is\r\n```\r\ntags:\r\n  - name: <2.1>\r\n    hash: bb0c6ae2f12a1b55df24ebce2067";
  355577929 rn: "Can we address one issue at a time. templating of the YAML file is one thing and";
  355579842 ijc: "You can already get the package hash from a git tag today already, it's currentl";
  355656944 w9n: "I made `pkgInfo` public and moved it into the config namespace I put GlobalConfi";
  355967540 ijc: "> I made pkgInfo public \r\n\r\nI don't like this, sorry. This struct should remain ";
  356205605 w9n: "@ijc I did what you suggested and exposed a `SetCommit` `GetHash` and `Image` fu";
  356250551 ijc: "You shouldn't need `SetCommit` since the only caller outside `pkglib` passes a h";
  356256042 w9n: "done";
  356556345 w9n: "so i rethought this and the whole templating problem might just not fit into lin";
  356567368 ijc: "I don't think there are objections to templating in general, my specific objecti";
  356571088 w9n: "If git tags or parameter like `foo=abcdef` are used than this would require e.g.";
  356572786 ijc: "We intend to release linuxkit as a whole, with a single consistent set of packag";
  356576666 rn: "I agree with @ijc that we do want some templating. We had several discussions in";
  356579465 justincormack: "I don't understand what the output of `linuxkit bake example/linuxkit_template.y";
  356581808 ijc: "It's a new file which you can pass to `linuxkit build` (after archiving it for r";
  356582385 justincormack: "Does it output to stdout or what? This needs to be documented properly.";
  356582717 w9n: "Okay, so I could live with this somewhere:\r\n```\r\ntags:\r\n    - name: v1.0\r\n      ";
  356585190 rn: "@w9n please read my comment