fix: route in-memory shuffle partitions to remote fetch when not present locally#17
Merged
Merged
Conversation
…ent locally When using in-memory shuffle (shuffle_location: memory), shuffle partitions are stored in the executor's local InMemoryShuffleManager. Previously, the shuffle reader incorrectly assumed all memory:// paths could be read locally. In distributed mode with multiple executors, a shuffle reader on executor A may need to read a partition that was written by executor B. The partition would have a memory:// path, but only exist in executor B's shuffle manager. This caused 'Shuffle partition not found in memory' errors. This fix modifies split_partition_locations() to check if a memory:// partition actually exists in the local shuffle manager before categorizing it for local read. If the partition doesn't exist locally, it's routed to the remote fetch path which uses the Arrow Flight service. The Flight service already correctly handles memory:// paths by reading from the executor's local shuffle manager. Fixes spiceai/spiceai#9290
phillipleblanc
approved these changes
Feb 5, 2026
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When using in-memory shuffle (shuffle_location: memory), shuffle partitions are stored in the executor's local InMemoryShuffleManager. Previously, the shuffle reader incorrectly assumed all memory:// paths could be read locally.
In distributed mode with multiple executors, a shuffle reader on executor A may need to read a partition that was written by executor B. The partition would have a memory:// path, but only exist in executor B's shuffle manager. This caused 'Shuffle partition not found in memory' errors.
This fix modifies split_partition_locations() to check if a memory:// partition actually exists in the local shuffle manager before categorizing it for local read. If the partition doesn't exist locally, it's routed to the remote fetch path which uses the Arrow Flight service. The Flight service already correctly handles memory:// paths by reading from the executor's local shuffle manager.
Fixes spiceai/spiceai#9290