make getProcessingDistributionsForWorkId threadsafe #29979
Conversation
…onditions with remove tracker
|
Assigning reviewers. If you would like to opt out of this review, comment R: @Abacn added as fallback since no labels match configuration Available commands:
The PR bot will only process comments in the main thread (not review comments). |
|
Thanks for explanation. For future reference, would you mind add a brief comment in code to indicate the need of synchronized? |
| @@ -117,7 +117,8 @@ public Optional<ActiveMessageMetadata> getActiveMessageMetadataForWorkId(String | |||
| return Optional.ofNullable(null); | |||
| } | |||
|
|
|||
| public Map<String, IntSummaryStatistics> getProcessingDistributionsForWorkId(String workId) { | |||
| public synchronized Map<String, IntSummaryStatistics> getProcessingDistributionsForWorkId( | |||
| String workId) { | |||
| if (!activeTrackersByWorkId.containsKey(workId)) { | |||
There was a problem hiding this comment.
| if (!activeTrackersByWorkId.containsKey(workId)) { | |
| // synchronized to avoid race condition of accessing activeTrackersByWorkId | |
| if (!activeTrackersByWorkId.containsKey(workId)) { |
| @@ -117,7 +117,8 @@ public Optional<ActiveMessageMetadata> getActiveMessageMetadataForWorkId(String | |||
| return Optional.ofNullable(null); | |||
| } | |||
|
|
|||
| public Map<String, IntSummaryStatistics> getProcessingDistributionsForWorkId(String workId) { | |||
| public synchronized Map<String, IntSummaryStatistics> getProcessingDistributionsForWorkId( | |||
There was a problem hiding this comment.
since removeTracker isn't synchronized it still seems like the race is possible.
An alternative since the activeTrackersByWorkId is a concurrent map would be to just call get() first and replace the containsKey check with comparing if null is returned.
similarly the completedProcessingMetrics.containsKey check could be replaced by using getOrDefault
There was a problem hiding this comment.
yeah i like that better. done!
|
CC: @clmccart @scwhittle this issue is marked as release blocker. Any update? |
clmccart
force-pushed
the
dofn-patch
branch
2 times, most recently
from
3be409d
to
a7e5cb0
Compare
clmccart
changed the title
* make getProcessingDistributionsForWorkId synchronized to avoid race conditions with remove tracker * correctly access the concurrent hashmap to avoid race conditions --------- Co-authored-by: Claire McCarthy <clairemccarthy@google.com>
make getProcessingDistributionsForWorkId threadsafe to avoid race conditions with remove tracker
the race condition happens in getProcessingDistributionsForWorkId because we check for the existence of the work id in the activeTrackersByWorkId map before retrieving the value from the map. if another thread removes the tracker from the map in-between us checking for its existence and actually grabbing the object, we will get a null tracker that then throws a NullPointerException when we try to access data on it.
Was able to repro this with a chaos test locally and confirmed that synchronizing the getProcessingDistributionsForWorkId method solves the issue. Decided to not commit the chaos test since it's a bit messy and deviates from test patterns in this repo.
Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:
addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, commentfixes #<ISSUE NUMBER>instead.CHANGES.mdwith noteworthy changes.See the Contributor Guide for more tips on how to make review process smoother.
To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md
GitHub Actions Tests Status (on master branch)
See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.