Skip to content

Conversation

@Matthew-Whitlock
Copy link
Collaborator

Fixing the weirdness w/ develop branch being rebase --no-ffed onto master

Matthew-Whitlock and others added 12 commits November 14, 2024 15:35
Add a recovery callback to destroy partially-recovered members
when interrupted by failure

Fix the case of one rank finishing a store where its partner
fails, followed by a commit on the succesfull rank. Now come to
a consensus on timestamps on group reinitialization.
…lize

More thought can be put in to this (e.g. if a rank has failed, but all remaining
ranks reach finalize, could we just finalize anyway?)
Possible when a rank fails after reaching finalize if the remaining ranks
inconsistently succeed/fail on the barrier finishing.
@Matthew-Whitlock Matthew-Whitlock merged commit 6e7b2d9 into develop Nov 22, 2024
2 checks passed
@Matthew-Whitlock Matthew-Whitlock deleted the main-to-dev branch November 22, 2024 18:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants