Fix race condition when we init lambda#746
Merged
Conversation
Collaborator
Author
|
Got some test problems, I'll force-push the fix. |
There's a couple of problems here: 1. We subscribe a telemetry endpoint before the service is ready. If we receive a flush to that endpoint, then we try to perform a flush and get an NPE. 2. The access to the flusher is unprotected across multiple goroutines. We just use an RWMutex for this, because it's rarely accessed and doesn't need anything fancy. 3. There's a race condition in the test because the nop write is (was) async.
tiedotguy
force-pushed
the
tdg/fix-lambda-race
branch
from
0e95a08 to
fb71920
Compare
Because we start the http server in the background, sometimes it's not ready when we try to call it.
tiedotguy
commented
Mar 18, 2025
|
|
||
| ch := &channeledHandler{ | ||
| chMaps: make(chan *gostatsd.MetricMap), | ||
| chMaps: make(chan *gostatsd.MetricMap, 1), |
Collaborator
Author
There was a problem hiding this comment.
We need this because the sendNop is now sync. If that doesn't go in to the channel, HttpForwarderHandlerV2.Run never starts using the virtual clock, and fixtures.NextStep() waits forever (because it's waiting for something to use the clock)
irisgve
reviewed
Mar 20, 2025
| func (hfh *HttpForwarderHandlerV2) Run(ctx context.Context) { | ||
| var wg wait.Group | ||
| wg.StartWithContext(ctx, hfh.sendNop) | ||
| hfh.sendNop(ctx) |
Member
There was a problem hiding this comment.
Just for my context, why do we need to send a nop before we do actual flushes?
Collaborator
Author
There was a problem hiding this comment.
The deepcheck checks if we've ever sent metrics - if we can't send metrics, we don't want to start.
We don't check how longs it's been though, since this is about misconfiguration, not backend going down.
Collaborator
Author
|
Actually, I guess I should include a changelog. |
irisgve
approved these changes
Mar 23, 2025
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
There's a couple of problems here:
We subscribe a telemetry endpoint before the service is ready. If we receive a flush to that endpoint, then we try to perform a flush and get an NPE.
The access to the flusher is unprotected across multiple goroutines. We just use an RWMutex for this, because it's rarely accessed and doesn't need anything fancy.
There's a race condition in the test because the nop write is (was) async.