fix(@libp2p/webrtc): delay closing datachannels #2048
Closed
+59
−4
Conversation
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
Tactical fix until we can update the spec to coordinate closing datachannels. Calling `.close` on a datachannel causes a stream reset - depending on the implementation there's a good chance any data waiting to be transmitted will be lost. As a workaround when closing a datachannel wait until it's `.bufferedAmount` property is `0`, then wait a bit longer before closing the channel. This is because although Firefox will have sent all the data by this point, Chrome may have not. This should be removed when the spec is updated.
maschad
reviewed
Sep 13, 2023
Pretty frequently, it turns out. It's what causes this kind of CI failure. |
Have you ran that test a few times on CI to ensure that this fix solves it? |
|
As discussed in triage, this may actually be an issue with Chrome, @achingbrain will file a report. Update: Report filed https://bugs.chromium.org/p/chromium/issues/detail?id=1484907 |
|
@achingbrainI I am re-opening this as it will unblock CI without us having to skip the test entirely, until https://bugs.chromium.org/p/chromium/issues/detail?id=1484907 is resolved. |
3 tasks
|
Closing in favour of #2073 |
achingbrain
added a commit
that referenced
this pull request
Oct 6, 2023
- Gracefully close streams on muxer shutdown - Refactor initiator/recipient flows for clarity - Wait for `bufferedAmount` to be `0` before closing a datachannel - Close datachannels on both initiator and recipient - Implements FIN_ACK for closing datachannels without data loss Supersedes #2048 --------- Co-authored-by: Chad Nehemiah <[email protected]>
achingbrain
added a commit
that referenced
this pull request
Oct 6, 2023
- Gracefully close streams on muxer shutdown - Refactor initiator/recipient flows for clarity - Wait for `bufferedAmount` to be `0` before closing a datachannel - Close datachannels on both initiator and recipient - Implements FIN_ACK for closing datachannels without data loss Supersedes #2048 --------- Co-authored-by: Chad Nehemiah <[email protected]>
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.
Tactical fix until we can update the spec to coordinate closing datachannels.
Calling
.closeon a datachannel causes a stream reset - depending on the implementation there's a good chance any data waiting to be transmitted will be lost.As a workaround when closing a datachannel wait until it's
.bufferedAmountproperty is0, then wait a bit longer before closing the channel.This is because although Firefox will have sent all the data by this point, Chrome may have not.
This should be removed when the spec is updated.
Refs: libp2p/specs#575