fix: start process subreaper at top level to avoid shutdown hangs #380
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
For example, if an exec command is running when you Ctrl-C the Pebble daemon, it hangs. It's hanging because the ServiceManager is stopping the reaper early in the process, before TaskRunner has had a chance to abort the exec tasks (aborting an exec task sends SIGKILL to its pid via the tomb and command context). Then when TaskRunner.Stop is called, it calls tomb.Kill and then tomb.Wait on each task (exec) tomb, and because the reaper's not running, the tomb.Wait hangs. So the fix is to move the reaper.Start and reaper.Stop to the top level (inside the "run" command, which is also used for "enter"), instead of putting them in servstate.NewManager and ServiceManager.Stop. After doing this, some of the tests also had to be modified to start and stop the reaper. Fixes canonical#163
flotter
reviewed
Mar 11, 2024
flotter
reviewed
Mar 11, 2024
flotter
reviewed
Mar 11, 2024
There was a problem hiding this comment.
@benhoyt I've created a issue for us to also start the reaper top level, as we have a different entrypoint, and with reaper init removed at the servstate level, we will have to do this, before we move our Pebble reference forward.
benhoyt
added a commit
that referenced
this pull request
Mar 21, 2024
PR #380 recently changed the way the reaper is stopped. If `runDaemon()` returns an error the `defer` method overwrites the `err` return value in `err = reaper.Stop()`. If `err` was already set to something else, this will in effect overwrite the non-`nil` `err` with either `nil` (if `reaper.Stop()` returns `nil`) or a different error (if `reaper.Stop()` returns a different error). Because this defer method is the first one in the function, it will get called last, so it's the one that will have `err` set on entry if the function returns a non-`nil` error. --------- Co-authored-by: Ben Hoyt <[email protected]>
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.
For example, if an exec command is running when you Ctrl-C the Pebble daemon, it hangs.
It's hanging because the ServiceManager is stopping the reaper early in the process, before TaskRunner has had a chance to abort the exec tasks (aborting an exec task sends SIGKILL to its pid via the tomb and command context).
Then when TaskRunner.Stop is called, it calls tomb.Kill and then tomb.Wait on each task (exec) tomb, and because the reaper's not running, the tomb.Wait hangs.
So the fix is to move the reaper.Start and reaper.Stop to the top level (inside the "run" command, which is also used for "enter"), instead of putting them in servstate.NewManager and ServiceManager.Stop.
After doing this, some of the tests also had to be modified to start and stop the reaper.
Fixes #163 and fixes #284