Ensure READY messages sent out by exiting worker are consumed prior to it actually exiting (issue #278) #279
Conversation
This file contains hidden or 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
This commit fixes a race-condition which can lead to `WorkerLostError: Worker exited prematurely: exitcode 155` and `WorkerLostError: Worker exited prematurely: exitcode 0` errors. The race-condition occurs when a message sent by `Worker.workloop` is not consumed by `ResultHandler` prior to the worker being cleaned up by `Pool._join_exited_workers`. The solution to the race-condition is to keep a count of messages `ResultHandler` receives from a worker, and for the worker to return only when this count matches its own counter of completed tasks. This ensures that a worker will only be cleaned up once `ResultHandler` and the `Worker` are in-sync, or after 30 seconds of waiting for them to be in-sync.
auvipy
requested changes
Apr 23, 2019
Member
auvipy
left a comment
There was a problem hiding this comment.
Hey, thanks for tackling this!! The code looks good so far. would you mind adding some tests for these changes? that would let the review easier.
Contributor
Author
|
Yep, definitely. Planning to add some tests during the weekend. |
Member
|
got the time to add test? |
Contributor
Author
Struggling to find the time to do so, but I might be able to free up a few hours this Sunday and get it done. |
The previous change did not account for "wait_for_job" raising a SystemExit exception. This change wraps the main processing logic of Worker.workloop in a try/finally block which ensures that _ensure_messages_consumed is always called.
2 tasks
auvipy
approved these changes
Jun 7, 2019
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
This commit fixes a race-condition which can lead to
WorkerLostError: Worker exited prematurely: exitcode 155andWorkerLostError: Worker exited prematurely: exitcode 0errors.The race-condition occurs when a message sent by
Worker.workloopis not consumed byResultHandlerprior to the worker being cleaned up byPool._join_exited_workers.The solution to the race-condition is to keep a count of messages
ResultHandlerreceives from a worker, and for the worker to return only when this count matches its own counter of completed tasks. This ensures that a worker will only be cleaned up onceResultHandlerand theWorkerare in-sync, or after 30 seconds of waiting for them to be in-sync.Fixes #278