Fix address terminated topic atomic updates (#4306) #4307
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
…mize the implementation to not take copies on Subscribe() or Unsubscribe().
Aaronontheweb
approved these changes
Mar 9, 2020
| @@ -35,7 +37,7 @@ public override AddressTerminatedTopic CreateExtension(ExtendedActorSystem syste | |||
| /// </summary> | |||
| internal sealed class AddressTerminatedTopic : IExtension | |||
| { | |||
| private readonly AtomicReference<HashSet<IActorRef>> _subscribers = new AtomicReference<HashSet<IActorRef>>(new HashSet<IActorRef>()); | |||
There was a problem hiding this comment.
Another ancient bug - looks like we tried to get too cute by half here. I agree that a lock should be a relatively simple fix for this.
Aaronontheweb
pushed a commit
to Aaronontheweb/akka.net
that referenced
this pull request
Mar 9, 2020
…mize the implementation to not take copies on Subscribe() or Unsubscribe(). (akkadotnet#4307)
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Outdated suggestions cannot be applied.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.
Submitting PR against v1.3, as that's what we're using and I was able to test the fix with.
The new implementation fixes both the race condition as well as optimizes the full copies of the _subscribers HashSet during Subscribe() and Unsubscribe(). The updates are protected with locks which makes the code simple and efficient. Publish() only takes a copy under lock to minimize the amount that the data is not available for other threads, which should be ok, as Publish() is only needed when a node is lost, so not too frequently.