Attempt to debug & fix TestReconnectBufferedUNIX for real #421
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
asf-stripe
force-pushed
the
asf-more-testreconnect-buffered-unix-debugging
branch
from
0567a66
to
6a9df00
Compare
asf-stripe
changed the title
Gerald Rule: Copy Observability on Veneur and Unilog pull requestscc @stripe/observability |
Turns out it was possible that a flush could be kicked off but could not reach the flushBackend, returning yet another ErrWouldBlock (which counted as an error, but broke assumptions further down). Now, we loop until both the outer and inner flush have succeeded, which should ensure this test succeeds on Travis.
asf-stripe
force-pushed
the
asf-more-testreconnect-buffered-unix-debugging
branch
from
6a9df00
to
3bc424b
Compare
|
R, @aditya-stripe? |
|
lgtm |
This was referenced Apr 3, 2018
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Outdated suggestions cannot be applied.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.
Summary
This PR hopefully 💯% fixes the TestReconnectBufferedUNIX for good. Double non-determinism in
selectdefault branches means we have to loop for the outer as well as the inner result.What happened was: Flushing a TraceBackend happens in two steps:
FlushAsynckicks off the flush in the background. If the goroutine is busy/not reading from the chan, this can returnErrWouldBlock. We handled this case!ErrWouldBlockif the backend is busy. We didn't handle this case!The part of the test that this was in was expecting a "real" error (not ErrWouldBlock) from the flush, and so in the relatively rare cases that ErrWouldBlock was returned, it would not have kicked off the flush and the later parts of the test would fail. Now, we handle both cases, so the test should correctly be in the right state always.
Motivation
This broke travis ci builds a bunch.
Test plan
Ran Travis CI tests a bunch
Rollout/monitoring/revert plan
Just merge, no changelog necessary.