feat: (do not merge) Block level partition shuffle rebased #19325
+115
−5
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
Add heuristic-based block-level shuffle for better load balancing when tables have few segments relative to cluster size. Changes: - Add BlockMod shuffle kind for block-level distribution - Add auto_block_shuffle_threshold setting (default=5, 0 to disable) - When segment_count < nodes * threshold, use block-level shuffle - Each executor filters blocks by block_idx % num_executors == executor_idx - Add info logging for shuffle strategy selection - Preserve partition kind during reshuffle to prevent data duplication
Move block_slot computation from executor-side (prune_segments_with_pipeline) to coordinator-side (redistribute_source_fragment). This ensures all executors use the same cluster view that was determined when the plan was created, preventing data duplication or loss if cluster membership changes. Changes: - Add block_slot field to DataSourcePlan - Compute block_slot in redistribute_source_fragment for BlockMod shuffle - Pass block_slot through plan instead of computing at execution time
…castWarehouse Block filtering is now controlled by plan.block_slot, not by partition kind. After reshuffle, all executors just process partitions sequentially. Also revert incorrect change in memory_table.rs (should use BroadcastCluster).
github-actions
bot
added
the
pr-feature
Contributor
Docker Image for PR
|
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Outdated suggestions cannot be applied.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.
I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/
Summary
perf only
Tests
Type of change
This change is