Merge lp:~gz/workspace-runner/ssh_server_alive into lp:workspace-runner
Proposed by
Martin Packman
Status: | Merged |
---|---|
Approved by: | Martin Packman |
Approved revision: | 27 |
Merged at revision: | 27 |
Proposed branch: | lp:~gz/workspace-runner/ssh_server_alive |
Merge into: | lp:workspace-runner |
Prerequisite: | lp:~gz/workspace-runner/sorted_walk_results |
Diff against target: |
78 lines (+8/-5) 2 files modified
workspace_runner/__init__.py (+1/-1) workspace_runner/tests/__init__.py (+7/-4) |
To merge this branch: | bzr merge lp:~gz/workspace-runner/ssh_server_alive |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Aaron Bentley (community) | Approve | ||
Review via email: mp+299722@code.launchpad.net |
Commit message
Add ServerAliveInte
Description of the change
It seems the timeouts on the race job under workspace runner are due to a long period with no output from the remote end and no activity from the runner.
The last line in the gating race output is the api/watcher suite. Next up would be apiserver, which takes >400 seconds to complete in the main race job.
ok github.
ok github.
We can try to work around this by getting the client to poke at a regular interval. For now, seeing if 2 minutes is good enough.
To post a comment you must log in.
Works for me.
I was a little surprised to see ssh_options becoming a class variable instead of an instance variable, but I guess it doesn't matter much-- class variables can always be overridden.