Join GitHub today
GitHub is home to over 20 million developers working together to host and review code, manage projects, and build software together.
Fix two agent wedgies. #8089
Conversation
| + // If the server is rate limiting, it will normally pause | ||
| + // before responding to the login request, but the pause is | ||
| + // in the realm of five to ten seconds. | ||
| + Timeout: time.Minute, |
wallyworld
Nov 16, 2017
Owner
Could we rename this to LoginTimeout?
So we have DialTimeout and LoginTimeout. Much clearer IMHO
howbazaar
Nov 16, 2017
Owner
Perhaps in another branch, but this change should be small in case we want to backport.
wallyworld
approved these changes
Nov 16, 2017
I think we can improve the select{} bits
|
$$merge$$ |
|
Status: merge request accepted. Url: http://ci.jujucharms.com/job/github-merge-juju |
jujubot
merged commit 93f6f62
into
juju:develop
Nov 16, 2017
1 check passed
continuous-integration/jenkins/pr-merge
This commit looks good
Details
howbazaar
deleted the
howbazaar:fix-agent-wedgies
branch
Nov 16, 2017
added a commit
that referenced
this pull request
Dec 13, 2017
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
howbazaar commentedNov 16, 2017
Fixes two places where agents can get wedged, which stops the agents from restarting.
First, the apicaller can get itself into a situation where it is waiting forever for a login response.
Secondly, the logsender can get itself stuck waiting for the websocket to be initialized.
This bug adds a timeout to the first, and handles the second by getting the LogSender in another goroutine and also waiting on the stop channel.
QA steps
Unit tests continue to work, and CI passes. Nothing spectacular.
Bug reference
https://bugs.launchpad.net/juju/+bug/1732588
https://bugs.launchpad.net/juju/+bug/1732587