Work-around for dask/distributed#7726 and dask-cuda#1148 #1151
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
The last atexit finalizer in the chain of finalizers that distributed registers needs to be one in the weakref module. Otherwise we observe hangs in the shutdown of UCX-Py where one thread has the gil but not the UCX spinlock and vice-versa. To ensure this, unregister the known shutdown finalizers in distributed, add our weakref finalizer, and then re-register them. Along with a final re-implementation of the "python shutting down" handler which must run before other handlers in distributed.
|
If @randerzander reports success, I think we could try this workaround? (cc @quasiben, @pentschev) |
|
Please see #1148 (comment), I think this may not be sufficient to fully resolve the issue without at least minor user intervention (assuming this problem is solely occurring in the client). |
|
Closing for now since it seems we don't have the necessity for this. |
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Outdated suggestions cannot be applied.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.
The last atexit finalizer in the chain of finalizers that distributed registers needs to be one in the weakref module. Otherwise we observe hangs in the shutdown of UCX-Py where one thread has the gil but not the UCX spinlock and vice-versa.
To ensure this, unregister the known shutdown finalizers in distributed, add our weakref finalizer, and then re-register them. Along with a final re-implementation of the "python shutting down" handler which must run before other handlers in distributed.