grpc-js: Fix connectivity state change event sequencing #2421
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
gnossen
approved these changes
Apr 12, 2023
| @@ -227,6 +227,8 @@ export class Subchannel { | |||
| } | |||
| const previousState = this.connectivityState; | |||
| this.connectivityState = newState; | |||
| process.nextTick(() => { | |||
There was a problem hiding this comment.
@murgatroid99
What's the purpose of this call?
It looks like some debugging leftover. It doesn't make anything asynchronous, it just adds an empty function to be called by Event Loop after synchronous part of the code is executed.
There was a problem hiding this comment.
You're right. That was part of a change that I was trying out before I went in a different direction.
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Outdated suggestions cannot be applied.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.
The test in #2419 revealed a couple of bugs related to the order of connectivity state change events:
PickFirstLoadBalancer#pickSubchannelinforms its parent of the state change before resolving its own internal state, and as a result arbitrary code can run at that point.Subchannel#transitionToStatecan eventually result in API calls that cause additional state changes, causing potential reentrancy intransitionToStatewhich can result in out-of-order state change notifications.The changes to fix this are:
PickFirstLoadBalancer#pickSubchannelto the end.listenerlist inSubchannelto a set to simplify interacting with it and to allow removing listeners during the iteration.transitionToStateare asynchronous with respect to surface API calls, to avoid reentranttransitionToStatecalls.I adapted the test from #2419 to test this combination of bugs.