Merge 2.7 on 2019-11-09. #10994
Closed
Merge 2.7 on 2019-11-09. #10994
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
It had been waiting for delay before it would exit, instead we can notice that we got the request to shutdown early. This didn't actually block any main threads, because it was only being triggered by the hub watcher. Still, it is cleaner, IMO.
When running a charm hook, we would notice if the hook exited nonzero, but if we failed to Start the hook, that error ended up being suppressed. This patch just puts the error in the same place as existing exit error.
…internal variable.
…54338 juju#10966 ## Please provide the following details to expedite Pull Request review: ### Checklist - [No] Checked if it requires a [pylibjuju](https://github.com/juju/python-libjuju) change? - [N/A] Added [integration tests](https://github.com/juju/juju/tree/develop/tests) for the PR? - [N/A ] Do comments answer the question of why design decisions were made? ## Description of change Don't suppress errors in Start. When running a charm hook, we would notice if the hook exited nonzero, but if we failed to Start the hook, that error ended up being suppressed. This patch just puts the error in the same place as existing exit error. ## QA steps You can write a charm that has an invalid 'install' hook. Something like: ```sh $ cd acceptancetests/repository/charms $ echo "exit 2" > dummy-sink/hooks/install $ juju deploy ./dummy-sink ``` Without this patch, it ignores the install hook failure, with this patch you get something like: ``` $ juju status ... Unit Workload Agent Machine Public address Ports Message dummy-sink/2* error idle 2 10.210.24.221 hook failed: "install" ``` ## Documentation changes None. ## Bug reference https://bugs.launchpad.net/juju/+bug/1854338
juju#10933 ## Description of change This is a small cleanup for the http dualPort code. If the http server was shutting down, the old code wouldn't notice until timeout was reached. This didn't actually block anything (as far as I can tell) because the dualListener was opened from a Hub message, which didn't block the other worker from shutting down. So all this really does is cleanup a goroutine a bit earlier. Still, since I noticed it, I figured it was worth landing. The test doesn't actually fail if you don't exit early, because we can't wait for the openAPIPort to return, because it is triggered from the Hub subscription. Though I think the setup of the test is still relevant (you shouldn't fail if requested to shut down before the timeout occurs.) ## QA steps New Unit test. You can see how it is triggered if you do 'go test -check.vv' where you can see that it notices right away that shutdown was requested. ## Documentation changes None. ## Bug reference None.
juju#10992 With some of the testing infrastructure coming under increasing load, it changes some of the timing for the tests. This is leading to some intermittent test failures where we didn't see them before. When tests that are full stack tests, like JujuConnSuite, and they create entities, we need to ensure that the internal events have been processed before we continue after creation, or we get spurious test failures like: ``` 09:13:48 unitupgrader_test.go:132: 09:13:48 // Initial event 09:13:48 wc.AssertOneChange() 09:13:48 /home/ubuntu/go/src/github.com/juju/juju/core/watcher/watchertest/notify.go:95: 09:13:48 c.Fatalf("watcher sent unexpected change: (_, %v)", ok) 09:13:48 ... Error: watcher sent unexpected change: (_, true) ```
…limiting juju#10987 Add configurable rate limiting for agent connections. This work took a sideways step as I needed to write tests for the rate limiting, but all the login tests were using JujuConnSuite. So I changed the login tests to instead be based on statetetesting.StateSuite, which is much lighter. ## QA steps ```sh juju bootstrap lxd test juju controller-config agent-ratelimit-max=5 juju controller-config agent-ratelimit-rate=500ms juju deploy ~jameinel/ubuntu-lite -n 5 for i in {1..5}; do juju add-unit ubuntu-lite -n 5 --to 0,1,2,3,4; done watch -c -n 1 juju status --color # in a different shell juju ssh -m controller 0 sudo service jujud-machine-0 restart # observe the reduced rate of reconnection in the status ``` ## Documentation changes Will need to document the two new controller config values.
juju#10968 ## Please provide the following details to expedite Pull Request review: ### Checklist - [x] Checked if it requires a [pylibjuju](https://github.com/juju/python-libjuju) change? _No API server change, only api CLI client._ - [ ]~~ Added [integration tests](https://github.com/juju/juju/tree/develop/tests) for the PR?~~ - [ ] ~~Added or updated [doc.go](https://discourse.jujucharms.com/t/readme-in-packages/451) related to packages changed?~~ - [x] Do comments answer the question of why design decisions were made? ---- ## Description of change Cloud facade has supported the ability to force update and force remove credentials. This PR catches up corresponding CLI. As a drive-by, I have renamed an internal variable in credential validity to maintain clarity. ## QA steps 1. bootstrap with credential A 2. add model with different credential B 3. remove credential B [will fail because a model uses this credential] 4. force remove credential [success] ## Bug reference Part 2 for https://bugs.launchpad.net/juju/+bug/1852412
# Conflicts: # apiserver/admin_test.go
wallyworld
approved these changes
Dec 9, 2019
|
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
f78517f