implement job allocation completions index enhancement #1072

fudali113 · 2019-05-27T07:20:41Z

No description provided.

k8s-ci-robot · 2019-05-27T07:20:42Z

Welcome @fudali113!

It looks like this is your first PR to kubernetes/enhancements 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes/enhancements has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. 😃

k8s-ci-robot · 2019-05-27T07:20:49Z

Hi @fudali113. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

k8s-ci-robot · 2019-05-27T07:20:59Z

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: fudali113
To fully approve this pull request, please assign additional approvers.
We suggest the following additional approver: mattfarina

If they are not already assigned, you can assign the PR to them by writing /assign @mattfarina in a comment when ready.

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

keps/sig-apps/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

fudali113 · 2019-05-30T14:37:16Z

/assign @mattfarina

k82cn · 2019-06-02T06:16:52Z

keps/sig-apps/20190523-support-completions-index-for-job.md

+## Summary
+
+[Job](https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/)
+is a Kubernetes CRD that help us run task, Job allows a user to specify the task or parallel task through either `podTemplate` and `completions`; but i use parallel task i must ensure task unrepeat in myself application, this is diffcult; if job support sahrds, i can use shards index to implement task unrepeat is so easy;


Can you give an example on what's kind of workload is running in the Pod, e.g. MPI, Spark? If it's pure customized application, why not crd + operator?

btw, we implemented similar feature by CRD+Operator at https://github.com/volcano-sh/volcano/blob/master/docs/design/job-api.md#plugins-for-job which maybe helpful for your case.

such as we have million user, each user has different user results, i need push one article to user where has some result。

we expected run these task in one platform，that platform must easy to use and in common use，so we choice k8s；and we don`t want to use more platform(because we need lower maintenance cost)

solution is more， but i expected i need be for maintenance do something is less and i see this feature is not implement in document. i think this feature can help more people easy to use k8s.(we company use argo to implement we case, but argo more feature not to use)

is there others have similar case with yours?

we company has this case and i think this is a common question. because this implement can make job application easy to scale.

but my issue not more people comment support this implement.

@k82cn

@fudali113 , IMO, that's a specific case, and prefer to handle that by CRD; anyway, approvers will give the finial decision :)

k82cn · 2019-06-02T06:18:56Z

keps/sig-apps/20190523-support-completions-index-for-job.md

+
+## Motivation
+
+I have a case need shards to job, i want to use kubernetes implement my case; i read the [document](https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/#parallel-jobs), this is a not implement function, so i want implement it;


hm... it's better to have a general motivation, e.g. resolve a problem that raised by several dev/user.

ok, i will addition it;

my motivation is i think shared index is a general feature in job platform, i want to easy to use job when user have relate demand and maintenance platform less

k82cn · 2019-06-02T06:22:06Z

keps/sig-apps/20190523-support-completions-index-for-job.md

+
+### Implementation Details/Notes/Constraints
+
+In the current implementation, the Job controller not allocation index on `.completions != nil`; in my implementation, when `.completions != nil` i will through completions value and succeeded pods get the available indexes, i will record index to the job create pods, and set env to the pod （user application get index is easy）; next i will allocation index to creating pod function, i will lock this step ensure not repeat index not more than one, if one index repeat, i will kill one on next this job created pod changes time;


How do we persist the index, e.g. annotation, new field of Pod? If not persist, how to recover it when kube-controller-manager failover or change leader?

BTW, please also highlight the interface in design doc for review, e.g. env name ?

ok, i will fix it;

i persist the index to labels, i choice label because i think easy to use on kubectl get po -l, label name is job-completions-index and i set JOB_COMPLETIONS_INDEX env in pod (user get this index in application easier to use )

k82cn · 2019-06-02T06:22:50Z

keps/sig-apps/20190523-support-completions-index-for-job.md

+
+### Risks and Mitigations
+
+The major risk with this change is the additional load on the apiserver since we need frequent each pod for get succeeded index and error index.


How many "additional load" are there? A ballpark should be ok :)

group with index on activePods

remove duplicate index activePods

allocation index need an lock

The above three points is additional operation，i think that additional load is acceptable. and i don`t know how to get how many additional load is better, can you help me ?

this is my first pr in k8s, maybe there's something wrong with it， look forward to your comments @k82cn

…itional load description

fudali113 · 2019-07-23T13:28:42Z

@k82cn @prydonius @kow3ns PTAL, thanks

fejta-bot · 2019-12-15T07:42:21Z

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

fejta-bot · 2020-01-14T08:30:03Z

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

palnabarun · 2020-01-14T16:55:46Z

/remove-lifecycle rotten

fengzixu · 2020-01-16T07:41:56Z

Is this enhancement still being handling? It seems like there is no response from the author for several months. If it is possible, I can re-create the enhancement like this and implement it.

@k82cn @palnabarun @kow3ns @janetkuo @fudali113

mattfarina · 2020-02-21T18:41:20Z

/unassign

mattfarina · 2020-02-21T18:42:46Z

@fudali113 adding to the agenda for the 2/24 SIG Apps meeting to discuss if you can make it.

fudali113 · 2020-03-19T06:30:57Z

@fudali113 adding to the agenda for the 2/24 SIG Apps meeting to discuss if you can make it.

sorry，i just saw this comment now；What can i do to continue this PR ? @mattfarina

fejta-bot · 2020-06-17T06:57:54Z

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

fejta-bot · 2020-07-17T07:38:08Z

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

fejta-bot · 2020-08-16T08:20:15Z

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

k8s-ci-robot · 2020-08-16T08:20:27Z

@fejta-bot: Closed this PR.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

fudali113 added 2 commits May 23, 2019 20:00

feat: 添加文档文件

6720da4

feat: add completions-index enhancement document

31e3972

k8s-ci-robot added do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels May 27, 2019

k8s-ci-robot added kind/kep Categorizes KEP tracking issues and PRs modifying the KEP directory sig/apps Categorizes an issue or PR as relevant to SIG Apps. size/M Denotes a PR that changes 30-99 lines, ignoring generated files. labels May 27, 2019

k8s-ci-robot requested review from kow3ns and prydonius May 27, 2019 07:21

fudali113 changed the title ~~[WIP] Add job enhancement~~ [WIP] implement job allocation completions index enhancement May 27, 2019

This was referenced May 27, 2019

implement job allocation completions index enhancement #1073

Closed

add completions index feature to Parallel Jobs kubernetes/kubernetes#66105

Closed

fix: change Graduation Criteria

a690260

fudali113 changed the title ~~[WIP] implement job allocation completions index enhancement~~ implement job allocation completions index enhancement May 27, 2019

k8s-ci-robot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label May 27, 2019

k8s-ci-robot assigned mattfarina May 30, 2019

k82cn reviewed Jun 2, 2019

View reviewed changes

fix: 1. add my case detail; 2. add persist index solution; 3. add add…

bd38fae

…itional load description

k8s-ci-robot added the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Dec 15, 2019

k8s-ci-robot added lifecycle/rotten Denotes an issue or PR that has aged beyond stale and will be auto-closed. and removed lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. labels Jan 14, 2020

k8s-ci-robot removed the lifecycle/rotten Denotes an issue or PR that has aged beyond stale and will be auto-closed. label Jan 14, 2020

k8s-ci-robot unassigned mattfarina Feb 21, 2020

fudali113 mentioned this pull request Mar 30, 2020

Reopen add completions index feature to Parallel Jobs kubernetes/kubernetes#89640

Closed

k8s-ci-robot added the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Jun 17, 2020

k8s-ci-robot added lifecycle/rotten Denotes an issue or PR that has aged beyond stale and will be auto-closed. and removed lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. labels Jul 17, 2020

k8s-ci-robot closed this Aug 16, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

implement job allocation completions index enhancement #1072

implement job allocation completions index enhancement #1072

fudali113 commented May 27, 2019

k8s-ci-robot commented May 27, 2019

k8s-ci-robot commented May 27, 2019

k8s-ci-robot commented May 27, 2019

fudali113 commented May 30, 2019

k82cn Jun 2, 2019 •

edited

Loading

k82cn Jun 2, 2019

fudali113 Jun 3, 2019

k82cn Jun 17, 2019

fudali113 Jul 5, 2019

k82cn Sep 16, 2019 •

edited

Loading

k82cn Jun 2, 2019

fudali113 Jun 3, 2019

k82cn Jun 2, 2019

fudali113 Jun 3, 2019

k82cn Jun 2, 2019 •

edited

Loading

fudali113 Jun 3, 2019

fudali113 commented Jul 23, 2019 •

edited

Loading

fejta-bot commented Dec 15, 2019

fejta-bot commented Jan 14, 2020

palnabarun commented Jan 14, 2020

fengzixu commented Jan 16, 2020 •

edited

Loading

mattfarina commented Feb 21, 2020

mattfarina commented Feb 21, 2020

fudali113 commented Mar 19, 2020 •

edited

Loading

fejta-bot commented Jun 17, 2020

fejta-bot commented Jul 17, 2020

fejta-bot commented Aug 16, 2020

k8s-ci-robot commented Aug 16, 2020


		## Motivation

		I have a case need shards to job, i want to use kubernetes implement my case; i read the [document](https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/#parallel-jobs), this is a not implement function, so i want implement it;


		### Implementation Details/Notes/Constraints

		In the current implementation, the Job controller not allocation index on `.completions != nil`; in my implementation, when `.completions != nil` i will through completions value and succeeded pods get the available indexes, i will record index to the job create pods, and set env to the pod （user application get index is easy）; next i will allocation index to creating pod function, i will lock this step ensure not repeat index not more than one, if one index repeat, i will kill one on next this job created pod changes time;


		### Risks and Mitigations

		The major risk with this change is the additional load on the apiserver since we need frequent each pod for get succeeded index and error index.

implement job allocation completions index enhancement #1072

implement job allocation completions index enhancement #1072

Conversation

fudali113 commented May 27, 2019

k8s-ci-robot commented May 27, 2019

k8s-ci-robot commented May 27, 2019

k8s-ci-robot commented May 27, 2019

fudali113 commented May 30, 2019

k82cn Jun 2, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

k82cn Sep 16, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

k82cn Jun 2, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

fudali113 commented Jul 23, 2019 • edited Loading

fejta-bot commented Dec 15, 2019

fejta-bot commented Jan 14, 2020

palnabarun commented Jan 14, 2020

fengzixu commented Jan 16, 2020 • edited Loading

mattfarina commented Feb 21, 2020

mattfarina commented Feb 21, 2020

fudali113 commented Mar 19, 2020 • edited Loading

fejta-bot commented Jun 17, 2020

fejta-bot commented Jul 17, 2020

fejta-bot commented Aug 16, 2020

k8s-ci-robot commented Aug 16, 2020

k82cn Jun 2, 2019 •

edited

Loading

k82cn Sep 16, 2019 •

edited

Loading

k82cn Jun 2, 2019 •

edited

Loading

fudali113 commented Jul 23, 2019 •

edited

Loading

fengzixu commented Jan 16, 2020 •

edited

Loading

fudali113 commented Mar 19, 2020 •

edited

Loading