parallelize(): fix occasional crashes on high load#904
Merged
alltilla merged 2 commits intoaxoflow:mainfrom Jan 15, 2026
Merged
parallelize(): fix occasional crashes on high load#904alltilla merged 2 commits intoaxoflow:mainfrom
parallelize(): fix occasional crashes on high load#904alltilla merged 2 commits intoaxoflow:mainfrom
Conversation
Calling log_msg_update_ack_and_ref_and_abort_and_suspended() and doing free/ack operations based on the returned value is atomic as long as the values passed to this function are not zero. The new acks_changed/refs_changed checks make sure that there is a real "non-zero to zero" transition before doing ack/free operations. The double-ack crash can be reproduced with parallelize(), using a big number of workers: workers(nproc) batch_size(100). In this specific case, a given message is going through 3 ref-cached threads: 1 producer and 2 consumers (LogScheduler, LogWriter). Signed-off-by: László Várady <[email protected]>
Signed-off-by: László Várady <[email protected]>
MrAnno
force-pushed
the
fix-refcache-ack-and-ref-and-rock-and-roll
branch
from
f5171eb to
fb90b56
Compare
bazsi
approved these changes
Jan 14, 2026
Member
bazsi
left a comment
bazsi
left a comment
There was a problem hiding this comment.
Great catch, and good solution. As a next step in calmer times, we should drop the whole of refcache.
alltilla
approved these changes
Jan 15, 2026
Member
|
Nice one! |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Calling log_msg_update_ack_and_ref_and_abort_and_suspended() and doing free/ack operations based on the returned value is atomic as long as the values passed to this function are not zero.
The new acks_changed/refs_changed checks make sure that there is a real "non-zero to zero" transition before doing ack/free operations.
The double-ack crash can be reproduced with parallelize(), using a big number of workers: workers(nproc) batch_size(100).
In this specific case, a given message is going through 3 ref-cached threads: 1 producer and 2 consumers (LogScheduler, LogWriter).