release-20.1: colexec: fix premature calls to IdempotentClose in hash router outputs #49624
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Outdated suggestions cannot be applied.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.
Backport 1/1 commits from #49333.
/cc @cockroachdb/release
Previously, we were passing in
toClose
slice of idempotent closers toevery outbox that is the output of a hash router. However, it is
possible that one stream will be closed before others, and this would
prompt the corresponding outbox to close all of the closers prematurely.
Other streams might still be active and ready to consume more data, but
that single outbox would close everything. This is now fixed by making
hash router responsible for closing the whole
toClose
slice, and itwill do so right before exiting
Run
method.Fixes: #49315.
Release note (bug fix): Previously, CockroachDB could return an internal
error or incorrect results on queries when they were run via the
vectorized execution engine and had a hash router in the DistSQL plan.
This could only occur with
vectorize=on
.