[hpc] Bound Ray plasma startup - #88
Merged
Merged
Conversation
Pass Ray’s plasma-socket connection-attempt limit through the shared Apptainer prefix. Large object-store mappings can exceed Ray’s ten-second default even when GCS is healthy, causing otherwise identical cluster starts to fail nondeterministically.
penfever
marked this pull request as ready for review
August 6, 2026 14:28
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Pass Ray’s plasma-socket connection-attempt limit through the shared Apptainer prefix while retaining the separate GCS-registration deadline. Large 40 GiB object-store mappings can exceed Ray 2.51.1’s ten-second default even when GCS is healthy, causing otherwise identical cluster starts to fail nondeterministically.
Jupiter job 1262020 created the plasma socket after about eight seconds. Job 1262271 exhausted all ten default attempts before the socket appeared and the head raylet aborted. Both used the same runtime and object-store size.
Jupiter job 1264144 then formed a four-GPU Ray cluster with a preallocated 40 GiB object store and completed 0:0 in 42 seconds using the new bound.