Revert "Revert AMM ReduceReplicas and parallel AMMs updates (#5335)" #5337
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
Collaborator
|
For the sake of not blocking more than what needs to, I'd suggest retiring this PR and going on with
which, put together, duplicate this PR - except that the first two can be merged in straight away. |
Member
Author
|
fine by me |
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Outdated suggestions cannot be applied.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 reverts the revert and adds another commit. The additional commit re-establishes a client+worker message send. This is misleading since the transition system on scheduler side switched to inplace dict mutations and the refactoring dropped these messages. I cannot reason yet how this leads up to our observed deadlock but it is a start.
For reference, I intended to remove this handler entirely in #5091 which outlines a subtle, yet non critical race condition. I believe it will become obsolete with #5046 as well. I will further investigate the code to look for more potential diffs but so far it look good
xref #5335 #5297