MCOL-3713: Join abort fix #1001
Merged
Conversation
This file contains 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
added 2 commits
January 6, 2020 18:03
Seems that segregateJoiners would double-lock a mutex. Not sure why we're only seeing it now. This is a checkpoint commit, probably not final for this bug.
A little cleanup. It's working.
Noticed a race on djsJoiners in abort() and segregateJoiners(). Re-added the mutex lock, but after risk of double lock grab.
drrtuy
reviewed
Jan 9, 2020
| djsJoiners.push_back(joiners[i]); | ||
| djsJoinerMap.push_back(i); | ||
| } | ||
| return; | ||
| } | ||
| #endif | ||
|
|
||
| boost::mutex::scoped_lock sl(djsLock); |
There was a problem hiding this comment.
the lock protects djs and djsJoiners. Abort() needs a consistent view of those, and it can come asynchronously.
jmrojas2332
approved these changes
Jan 15, 2020
This was referenced Jan 21, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Noticed a couple problems. 1) segregateJoiners() would double-lock djsMutex if any join exceeded the mem limit; 2) the logic to detect the mem overflow might have flagged joins as disk-joins even if that were disabled, requiring a longer than necessary abort process. Also it was fairly unreadble; cleaned it up & added a little inline documentation.
To verify I set the session var columnstore_um_mem_limit first between the PM and UM join limits and executed a large join, then set it below the PM join limit. Verified with gdb that no threads were orphaned; verified with valgrind & top that no memory was leaked. Enabled disk join and verified the same.