Do not share Weight between Drivers #133446
Merged
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
|
Hi @dnhatn, I've created a changelog YAML for you. |
elasticsearchmachine
added
the
Team:Analytics
|
Pinging @elastic/es-analytical-engine (Team:Analytics) |
nik9000
reviewed
Sep 3, 2025
| } | ||
| } | ||
|
|
||
| private static class OwningWeight extends FilterWeight { |
There was a problem hiding this comment.
Could we put the Query into LuceneScorer instead of into the Weight? There's already a bunch of useful stuff. We could do:
+ private final Query query;
private final Weight weight;
and check there.
There was a problem hiding this comment.
Sure, I've updated the LuceneScorer to keep both the query and tags.
nik9000
approved these changes
Sep 3, 2025
|
Makes sense to me. |
|
Thanks Nik! |
rjernst
pushed a commit
to rjernst/elasticsearch
that referenced
this pull request
Sep 9, 2025
We have encountered the following error in serverless: ``` java.lang.NullPointerException: Cannot invoke \"org.apache.lucene.search.BulkScorer.score(org.apache.lucene.search.LeafCollector, org.apache.lucene.util.Bits, int, int)\" because \"this.bulkScorer\" is null at org.elasticsearch.compute.lucene.LuceneOperator$LuceneScorer.scoreNextRange(LuceneOperator.java:233) at org.elasticsearch.compute.lucene.LuceneSourceOperator.getCheckedOutput(LuceneSourceOperator.java:307) at org.elasticsearch.compute.lucene.LuceneOperator.getOutput(LuceneOperator.java:143) at org.elasticsearch.compute.operator.Driver.runSingleLoopIteration(Driver.java:272) at org.elasticsearch.compute.operator.Driver.run(Driver.java:186) at org.elasticsearch.compute.operator.Driver$1.doRun(Driver.java:420) ``` I spent considerable time trying to reproduce this issue but was unsuccessful, although I understand how it could occur. Weight should not be shared between threads. Most Weight implementations are safe to share, but those for term queries (e.g., TermQuery, multi-term queries) are not, as they contain mutable This change proposes to stop sharing Weight between Drivers.
dnhatn
added a commit
to dnhatn/elasticsearch
that referenced
this pull request
Sep 10, 2025
This reverts commit 2f10065. # Conflicts: # x-pack/plugin/esql/compute/src/main/java/org/elasticsearch/compute/lucene/TimeSeriesSourceOperator.java
dnhatn
added a commit
that referenced
this pull request
Sep 10, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
We have encountered the following error in serverless:
I spent considerable time trying to reproduce this issue but was unsuccessful, although I understand how it could occur. Weight should not be shared between threads. Most Weight implementations are safe to share, but those for term queries (e.g., TermQuery, multi-term queries) are not, as they contain mutable TermStates.
This change proposes to stop sharing Weight between Drivers.
I am not sure if we should backport this to 8.19 and 9.1, since without the SliceQueue improvement from #132774, this fix may slow down queries due to the increased cost of creating more Weight instances.