[fix][prometheusexporter] Race condition between start and shutdown #36164
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
…pen-telemetry#36139) Signed-off-by: Argannor <[email protected]>
Argannor
force-pushed
the
fix/prometheusexporter-shutdown-server-2
branch
from
3c9ef63 to
83a5851
Compare
Signed-off-by: Argannor <[email protected]>
pjanotti
approved these changes
Nov 8, 2024
dmitryax
approved these changes
Nov 8, 2024
pull bot
pushed a commit
to abaguas/opentelemetry-collector-contrib
that referenced
this pull request
Nov 9, 2024
…pen-telemetry#36164) #### Description Adjusted the Start and Shutdown sequence of the prometheusexporter to prevent a race condition causing the `close tcp 127.0.0.1:8999: use of closed network connection` error as observed in open-telemetry#36139. The behaviour was changed in the following ways: - If an error occurs during the creation of the server, close the listener immediately, leaving shutdown a noop - Since `srv.Shutdown` will close all open listeners, the explicit call to `ln.Close` in the shutdown routine was removed #### Link to tracking issue Fixes open-telemetry#36139 #### Testing Unit testing, I temporarily increased the number of iterations on `TestPrometheusExporter` to 2000. The error did no longer occur. However sporadically another error occured: ``` === RUN TestPrometheusExporter prometheus_test.go:103: Error Trace: C:/development/code/opentelemetry-collector-contrib/exporter/prometheusexporter/prometheus_test.go:84 Error: Received unexpected error: listen tcp 127.0.0.1:8999: bind: Only one usage of each socket address (protocol/network address/port) is normally permitted. Test: TestPrometheusExporter --- FAIL: TestPrometheusExporter (1.16s) ``` I assume that this is because the OS (in my case Windows) won't always close the underlying sockets immediately, blocking it for some time afterwards. I'm not sure there is anything we can do about that. --------- Signed-off-by: Argannor <[email protected]>
sbylica-splunk
pushed a commit
to sbylica-splunk/opentelemetry-collector-contrib
that referenced
this pull request
Dec 17, 2024
…pen-telemetry#36164) #### Description Adjusted the Start and Shutdown sequence of the prometheusexporter to prevent a race condition causing the `close tcp 127.0.0.1:8999: use of closed network connection` error as observed in open-telemetry#36139. The behaviour was changed in the following ways: - If an error occurs during the creation of the server, close the listener immediately, leaving shutdown a noop - Since `srv.Shutdown` will close all open listeners, the explicit call to `ln.Close` in the shutdown routine was removed #### Link to tracking issue Fixes open-telemetry#36139 #### Testing Unit testing, I temporarily increased the number of iterations on `TestPrometheusExporter` to 2000. The error did no longer occur. However sporadically another error occured: ``` === RUN TestPrometheusExporter prometheus_test.go:103: Error Trace: C:/development/code/opentelemetry-collector-contrib/exporter/prometheusexporter/prometheus_test.go:84 Error: Received unexpected error: listen tcp 127.0.0.1:8999: bind: Only one usage of each socket address (protocol/network address/port) is normally permitted. Test: TestPrometheusExporter --- FAIL: TestPrometheusExporter (1.16s) ``` I assume that this is because the OS (in my case Windows) won't always close the underlying sockets immediately, blocking it for some time afterwards. I'm not sure there is anything we can do about that. --------- Signed-off-by: Argannor <[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.
Description
Adjusted the Start and Shutdown sequence of the prometheusexporter to prevent a race condition causing the
close tcp 127.0.0.1:8999: use of closed network connectionerror as observed in #36139.The behaviour was changed in the following ways:
srv.Shutdownwill close all open listeners, the explicit call toln.Closein the shutdown routine was removedLink to tracking issue
Fixes #36139
Testing
Unit testing, I temporarily increased the number of iterations on
TestPrometheusExporterto 2000. The error did no longer occur. However sporadically another error occured:I assume that this is because the OS (in my case Windows) won't always close the underlying sockets immediately, blocking it for some time afterwards. I'm not sure there is anything we can do about that.
Documentation
n/a