-
Notifications
You must be signed in to change notification settings - Fork 579
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
redpanda: node decommission test failure #2142
Comments
I am looking at this issue, seem like a false positive.
What is interesting tho is the fact that consume failed to commit offsets:
|
Related: #2388 is a failure on the same test, but looks to be a different failure mode. @mmaslankaprv when you say this is a false positive, do you mean the test should be doing something different? The default poll interval threshold in kafka client is 5 mins (I'm not aware of the test modifying it) so it doesn't seem that likely that this was just a race |
@dswang No this shouldn't block: it's unlikely to be a regression in 21.10. |
Believed to be stable now, explanation in ticket: redpanda-data#2142 Signed-off-by: John Spray <jcs@vectorized.io>
Fixed mapping of raft error code to Kafka error codes in offset commit handling code path. When offset commit request fails we need to preserve the error semantic from Kafka broker. All the error indicating that current node is not a leader should be translated to `NOT_COORDINATOR` error. Other errors should be translated to generic server error as the client should rebuild its state instead of assuming particular broker behavior. Fixes: redpanda-data#2142 Signed-off-by: Michal Maslanka <michal@vectorized.io>
Fixed mapping of raft error code to Kafka error codes in offset commit handling code path. When offset commit request fails we need to preserve the error semantic from Kafka broker. All the error indicating that current node is not a leader should be translated to `NOT_COORDINATOR` error. Other errors should be translated to generic server error as the client should rebuild its state instead of assuming particular broker behavior. Fixes: redpanda-data#2142 Signed-off-by: Michal Maslanka <michal@vectorized.io>
Fixed mapping of raft error code to Kafka error codes in offset commit handling code path. When offset commit request fails we need to preserve the error semantic from Kafka broker. All the error indicating that current node is not a leader should be translated to `NOT_COORDINATOR` error. Other errors should be translated to generic server error as the client should rebuild its state instead of assuming particular broker behavior. Fixes: redpanda-data#2142 Signed-off-by: Michal Maslanka <michal@vectorized.io> (cherry picked from commit 07ee7c5)
https://buildkite.com/vectorized/redpanda/builds/1723
The text was updated successfully, but these errors were encountered: