/juju

merge master to lxd-container-type #4396

Merged
merged 421 commits into from Feb 12, 2016

Conversation

Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
10 participants
@tych0 @jujubot @dooferlad @tasdomas @frankban @perrito666 @wallyworld @davecheney @bogdanteleaga @sinzui
@tych0
Contributor

tych0 commented Feb 11, 2016

(Review request: http://reviews.vapour.ws/r/3838/)

Dimiter Naydenov and others added some commits Jan 5, 2016

Moved common helpers to providerSuite
eb7ad5e
Simplified a couple of places to use cached API capabilities
cd4abf5
Extra whitespace dropped
0ec2cd0
Moved legacy NetworkInterfaces()-related tests to interfaces_test.go
a785ee0
Implemented NetworkInterfaces() and a simple test for it, using the n… 
…ew API
3082395
Added NewAddressOnSpace and a test for it
98e6084
Added NewAddressesOnSpace as well, improved tests
85d6750
Added SpaceName at the end of the address string representation, if n… 
…on-empty
f2af612
Finished new API NetworkInterfaces() implementation and tests
c1bafc2
StartInstance uses the new API to get the InterfaceInfo if possible
ff6e36f
Fixed a go vet warning
29fd0a8
Added LKK cards for TODOs, added MTU to network.InterfaceInfo
4ee89a6
@jujubot
Merge pull request #4043 from dimitern/maas-spaces-better-interfaces 
provider/maas: Better NetworkInterfaces() using the new API

Now the full network interface details for an instance are extracted using the
new 1.9+ MAAS API, when we detect it is available. The new implementation allows
us to pave the way for multi-NIC containers and is also simpler than the legacy
approach (which is still supported on older versions of MAAS).

Also tested live on 1.9 in addition to the unit tests.

(Review request: http://reviews.vapour.ws/r/3462/)
bad67d0
Basic test infrastructure plus apiserver improvement
babb002
Remove unneeded EnvironObserver stuff from the worker and add Environ… 
…Config to the apiserver
8369070
Really add EnvironConfig to the API
601a4bf
More resilient handling of NICs w/o IPs and more logging
b99dfad
Fixed a go vet issue
eac67ab
Test EnvironConfig
dbfc220
Remove EnvironWatcher from api
d0bd80d
Skeleton test
c6f2586
Test fix
38db4e9
@jujubot
Merge pull request #4046 from dimitern/maas-spaces-demo-hotfix 
maas: Better handling of NICs with no IPs and more logging

Interfaces with no IP addresses assigned are now skipped and logged rather than causing errors. Before trying to parse the interfaces from the node details, we always refresh them to ensure we have the latest details. Added more logs around the uniter NetworkConfig API used by network-get. NICs configured as "auto" or "unconfigured" are handled better, as we read the interfaces after startNode succeeds, rather than after acquireNode (where "auto" NICs don't have IPs yet). So "static"-only NICs are no longer a requirement.

Tested live on 1.9rc4 with both bindings and storage to ensure the storage code wasn't broken.
Confirmed to work for the ceph demo by David.
 
Review request: http://reviews.vapour.ws/r/3465/)
67762db
Test ListSubnets
8f53f43
Test extension
4d8a1f7
@dooferlad
Add endpoint binding support to charm bundle deployment.
ba78137
Indentation arguments
c441990
@jujubot
Merge pull request #4056 from dooferlad/bundle-endpoint-bindings 
Add endpoint binding support to charm bundle deployment.

Adds support for reading endpoint bindings from a charm bundle, for example:
```yaml
services:
    wordpress:
        charm: wordpress
        num_units: 1
        bindings:
            db: db
            url: public
    mysql:
        charm: mysql
        num_units: 1
        bindings:
            server: db
relations:
    - ["wordpress:db", "mysql:server"]
```
There is no support for overriding endpoint bindings on the command line like there is for storage, but it could, in theory, be added after the syntax has been agreed.

(Review request: http://reviews.vapour.ws/r/3475/)
d829f9d
Add package_test
28f5e0f
Initial test skeleton
5ea5933
Minor fixes
77400b3
@jujubot
Merge pull request #4054 from voidspace/maas-spaces-networking-api8 
Tests for discoverspaces apiserver

Tests for discoverspaces apiserver. EnvironWatcher is no longer used in the worker, api client or apiserver. EnvironConfig is implemented directly instead.

(Review request: http://reviews.vapour.ws/r/3473/)
6220a2b
Comment update
d1d03bc
Properly checking error results
3dde46b
provider/maas: DNS entries do not demarcate stanzas 
When parsing an interfaces file, entries beginning with 'dns-' do not
demarcate stanzas. DNS entries are now considered part of the iface
stanza that is being parsed.

For example, the following interfaces file:

   auto eth0
   iface eth0 inet dhcp

   dns-nameservers 10.10.19.2
   dns-search maas-19

we would previously bridge and render as:

   iface eth0 inet manual

   auto br-eth0
   iface br-eth0 inet dhcp
       bridge_ports eth0

   dns-nameservers 10.10.19.2
   dns-search maas-19

but is now parsed and rendered as:

   iface eth0 inet manual

   auto br-eth0
   iface br-eth0 inet dhcp
       dns-nameservers 10.10.19.2
       dns-search maas-19
       bridge_ports eth0

Fixes [LP:#1532176](https://bugs.launchpad.net/juju-core/+bug/1532176)
8568630
Add additional test from dellstack setup
84e785d
@jujubot
Merge pull request #4062 from frobware/maas-spaces-lp1516891 
provider/maas: DNS entries do not demarcate stanzas

When parsing an interfaces file, entries beginning with 'dns-' do not
demarcate stanzas. DNS entries are now considered part of the iface
stanza that is being parsed.

For example, the following interfaces file:

   auto eth0
   iface eth0 inet dhcp

   dns-nameservers 10.10.19.2
   dns-search maas-19

we would previously bridge and render as:

   iface eth0 inet manual

   auto br-eth0
   iface br-eth0 inet dhcp
       bridge_ports eth0

   dns-nameservers 10.10.19.2
   dns-search maas-19

but is now parsed and rendered as:

   iface eth0 inet manual

   auto br-eth0
   iface br-eth0 inet dhcp
       dns-nameservers 10.10.19.2
       dns-search maas-19
       bridge_ports eth0

Fixes [LP:#1532176](https://bugs.launchpad.net/juju-core/+bug/1532176)

(Review request: http://reviews.vapour.ws/r/3481/)
92d040b
Merged ipAddresses() into legacyAddresses() as it's only used there a… 
…nyway
fb3add1
Simplified instance.Addresses() to use maasObjectNetworkInterfaces in… 
…ternally if possible
668ac10
Cleaning up obsolete/legacy/no longer needed code paths
b0e9f8b
Name transformation
66bd8f9
Improve space name wrangling
70a5cc4
Comments and error message
f9c4fa0
Added TODO
f35f550
Comment update
059542a
Adding export_test
744739d
New tests
62d5b9e
Merge remote-tracking branch 'upstream/maas-spaces' into maas-spaces-… 
…networking-api9
50c1d1c
Test and error handling update
a972f52
Test worker is not a finished worker
7dad603
Update dummy provider with configurable space discovery support
7ae19fc
Support broken spaces on dummy provider
2b06689
Adding spaces to dummy provider and test extension
b1500f2
Initial extended support for SpaceDiscovery in dummy provider Subnets… 
… method
e888f4a
Fixes
909e5bd
Changes after second review
fd4f137
Update dummy provider subnets
e625c03
Call convertSpaceName unconditionally to catch duplicates
29bb39f
Wait for spaces to be created in test
ff66a07
Add a wait in test
b29dc22

Michael Foord and others added some commits Feb 9, 2016

Added SpaceProviderId to String output
734b9fe
@jujubot
Merge pull request #4352 from voidspace/maas-spaces-address-provider-id 
Add SpaceProviderId to network.Address

Precursor to using space provider id instead of name in MAAS provider.

(Review request: http://reviews.vapour.ws/r/3796/)
d036eef
apiserver/provisioner: Simplified a couple of tests, extracted Provis… 
…ioningInfo code and tests in separate files
2976629
juju/testing: added AddTestingServiceWithBindings, dropped AddTesting… 
…ServiceWithNetworks
e874d88
apiserver/provisioner: Added minimal test for ProvisioningInfo with E… 
…ndpointBindings
3dca404
state: Changed bindings tests to accept empty spaces as valid and mis… 
…sing bindings as defaults
6e3ff05
state: All assumptions about endpoint bindings changed in tests
162979e
state: Endpoint bindings now handled as expected in tests
e04dab3
apiserver/provisioner: no more network.DefaultSpace for bindings
fb922d5
apiserver/service: no more network.DefaultSpace in tests for unspecif… 
…ied bindings
4edab94
juju: DeployService populates incompletely specified bindings
b6c3c3b
cmd/juju/service: Added tests for deploying charm with bindings (not … 
…just bundle)
4f6681f
api[server]/uniter: Use machine private address for unspecified endpo… 
…ints; added more tests
e641f72
apiserver/uniter: fixed a go vet issue
fae506d
service/windows: Fixed a go vet warning
095df83
apiserver/provisioner: Added bugs/LKK card links to a few TODOs
6b964c8
@jujubot
Merge pull request #4331 from dimitern/maas-spaces-optional-bindings 
Make service endpoint bindings optional and drop network.DefaultSpace dependency

RB link: http://reviews.vapour.ws/r/3773/
63dc698
@davecheney
rcp: simplify serverError type 
serverError was over compliated. It did not need to inherit from
requestError, and the way it did was overengineered. serverError just
needs to extend the error interface to match the ErrorCoder interface,
and as serverError only knows one status code, we can hard code that as
well.
863b52d
@jujubot
Merge pull request #4357 from davecheney/rpc-simplify-server-error 
rcp: simplify serverError type

serverError was over compliated. It did not need to inherit from
requestError, and the way it did was overengineered. serverError just
needs to extend the error interface to match the ErrorCoder interface,
and as serverError only knows one status code, we can hard code that as
well.

(Review request: http://reviews.vapour.ws/r/3802/)
2c11d4b
@davecheney
rpc: use errors.Annotate 
Use errors.Annotate to include the "request error: " boilerplate that
other code (almost always tests) expects. This lets callers use
errors.Cause to unwrap the error to recover the *rpc.RequestError type
without having to restort to string inspection of the .Error() value.
e098f17
Merge branch 'master' into maas-spaces-merge-master-2c11d4b
fe38916
@jujubot
Merge pull request #4367 from dimitern/maas-spaces-merge-master-2c11d4b 
Merge master @ 2c11d4b into maas-spaces
9305acc
state: Refactored and extended spaces tests to set expected behavior
d5ff47b
state: Space.ProviderId now unique per model (when set)
d7cd4c2
state: Fixed EnsureDead, Remove, and Refresh for spaces
100ed7f
state: Fixed an import in spaces and dropped a TODO in tests
88178e5
state: Refactored and simplified Subnet tests, while adding more case… 
…s for the expected behavior
0fd5b56
state: EnsureDead and Remove for subnets fixed to match test expectat… 
…ions
c994bc5
state: Fixed a few typos in spaces tests
08c5dcc
state: Moved NewStateForModelNamed to ConnSuite to allow its reuse
f06d9a1
state: Added cross-model tests for subnets, similar to those for spaces
65830a5
state: Subnet IDs and ProviderId now unique per model, like for spaces
ed2fd14
@jujubot
Merge pull request #4354 from dimitern/maas-spaces-state-fix-spaces-s… 
…ubnets

Allow state.Space and Subnet with duplicate ProviderIds across models

Needed to allow discoverspaces to work in models other than the
controller model.

Also includes some improvements to tests.

(Review request: http://reviews.vapour.ws/r/3798/)
4dba9fb
lp1543216 - Use consistent addr for replicaset. 
The same logic should be used to select an address
for the mongo server during bootstrap and after
the system comes up.

After bootstrap, addresses for a machine are run
though MergedAddresses, which de-duplicates
provider addresses.  We should do the same during
bootstrap to avoid churning mongo by changing the
address after bootstrap completes.
ce8bc5a
@jujubot
Merge pull request #4373 from cherylj/1543216-set-right-mongo-addr-at… 
…-bootstrap

lp1543216 - Use consistent addr for replicaset.

The same logic should be used to select an address
for the mongo server during bootstrap and after
the system comes up.

After bootstrap, addresses for a machine are run
though MergedAddresses, which de-duplicates
provider addresses.  We should do the same during
bootstrap to avoid churning mongo by changing the
address after bootstrap completes.

(Review request: http://reviews.vapour.ws/r/3816/)
99a4e07
@bogdanteleaga
environs/config: add automatically-retry-hooks option
b400552
@bogdanteleaga
Make automatically-retry-hooks non immutable
a989490
@jujubot
Merge pull request #4366 from davecheney/rpc-use-errors-annotate 
rpc: use errors.Annotate

Use errors.Annotate to include the "request error: " boilerplate that
other code (almost always tests) expects. This lets callers use
errors.Cause to unwrap the error to recover the *rpc.RequestError type
without having to restort to string inspection of the .Error() value.

(Review request: http://reviews.vapour.ws/r/3809/)
b4e2785
@sinzui
Increment juju to 2.0-beta1
5be08fe
@jujubot
Merge pull request #4378 from sinzui/inc-2.0-beta1 
Increment juju to 2.0-beta1
c215028
Merge branch 'master' into maas-spaces-merge-master-c215028
30a808f
@jujubot
Merge pull request #4386 from dimitern/maas-spaces-merge-master-c215028 
Merge master @ c215028 into maas-spaces

Hopefully for the last time..

(Review request: http://reviews.vapour.ws/r/3829/)
f4de48d
@davecheney
apiserver/params: remove ClientError helper 
Updates LP#1538583

api/apiclient.APICall was the sole caller of params.ClientError.
params.Client error would rewrite the error value iff it was
rpc.RequestError, claiming that it was leaking unnecessary information
to the caller, specfically the "request error: " prefix. In b4e2785
the error prefix was convered to an errors.Annotate wrapper removing
the requirement to convert the error type.

Now, by removing params.ClientError, we can also remove the
errors.Annotate wrapper as the _type of the error_ tells us what it is.

So, callers get

a. A nice annotate error backtrace using errors.Trace
b. Access to the original *rpc.RequestError to inspect if they want.
e11ae23
@jujubot
Merge pull request #4361 from davecheney/apiserver-params-remove-clie… 
…nterror

apiserver/params: remove ClientError helper

Updates LP#1538583

api/apiclient.APICall was the sole caller of params.ClientError.
params.Client error would rewrite the error value iff it was
rpc.RequestError, claiming that it was leaking unnecessary information
to the caller, specfically the "request error: " prefix. In b4e2785
the error prefix was convered to an errors.Annotate wrapper removing
the requirement to convert the error type.

Now, by removing params.ClientError, we can also remove the
errors.Annotate wrapper as the _type of the error_ tells us what it is.

So, callers get

a. A nice annotated backtrace using errors.Trace
b. Access to the original *rpc.RequestError to inspect if they want.
f1ca8c7
@jujubot
Merge pull request #4392 from juju/maas-spaces 
Merge maas-spaces into master

We had a Blessed build for both master and maas-spaces and last few CI runs seem to indicate similar reliability of both branches, I think it's time to merge now.

(Review request: http://reviews.vapour.ws/r/3834/)
4ccd006
@jujubot
Merge pull request #4227 from bogdanteleaga/config-add-retry-bool 
environs/config: add automatically-retry-hooks option

What the title says.

If I missed anything please let me know.

(Review request: http://reviews.vapour.ws/r/3665/)
a49dab5
Merge branch 'master' into lxd-container-type 
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
19844cc
un-fuck tools/lxdclient after master merge 
For whatever reason git couldn't figure this out, so this is basically a
cherry-pick of 2a17adb on top of this
branch.

Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
59082d9
more un-fucking of master merge 
This is a cherry pick of 70c8b25 and
f8e3c5e.

Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
c42a58f
@tych0
Contributor

tych0 commented Feb 11, 2016

$$merge$$
@jujubot
Contributor

jujubot commented Feb 11, 2016

Status: merge request accepted. Url: http://juju-ci.vapour.ws:8080/job/github-merge-juju
@jujubot
Contributor

jujubot commented Feb 11, 2016

Build failed: Tests failed
build url: http://juju-ci.vapour.ws:8080/job/github-merge-juju/6362
@tych0
Contributor

tych0 commented Feb 11, 2016

$$merge$$
@jujubot
Contributor

jujubot commented Feb 11, 2016

Status: merge request accepted. Url: http://juju-ci.vapour.ws:8080/job/github-merge-juju
@jujubot
Contributor

jujubot commented Feb 12, 2016

Build failed: Tests failed
build url: http://juju-ci.vapour.ws:8080/job/github-merge-juju/6363
remove unused import 
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
790522a
@tych0
Contributor

tych0 commented Feb 12, 2016

$$merge$$
@jujubot
Contributor

jujubot commented Feb 12, 2016

Status: merge request accepted. Url: http://juju-ci.vapour.ws:8080/job/github-merge-juju

jujubot added a commit that referenced this pull request Feb 12, 2016

@jujubot
Merge pull request #4396 from tych0/lxd-container-type 
merge master to lxd-container-type



(Review request: http://reviews.vapour.ws/r/3838/)
7694111

@jujubot jujubot merged commit 7694111 into juju:lxd-container-type Feb 12, 2016

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment