-
Notifications
You must be signed in to change notification settings - Fork 548
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
CI Failure (key symptom) in ShadowIndexingWhileBusyTest.test_create_or_delete_topics_while_busy
#18454
Comments
The errors appear quite close to a truncation triggered by gc
|
The condition https://github.com/redpanda-data/redpanda/blob/dev/src/v/cloud_storage/remote_partition.cc#L581-L588 should prevent the error because the searched for offset is now below the log start, but it does not, maybe the offset returned by |
debugging in #18465 - seems to only happen on CDT and arm64 |
The sequence of events appears to be: Offset is advanced ahead of all segments in manifest to satisfy retention (5204 appears to be greater than all segments in manifest)
During query the log start offset is now computed as nullopt. We cannot compute the kafka offset for a log offset outside range:
Because the log start offset is nullopt, the condition to avoid error is missed https://github.com/redpanda-data/redpanda/blob/dev/src/v/cloud_storage/remote_partition.cc#L581-L589 |
We should possibly check for a full log truncation coinciding with a null kafka start offset, and if the query then falls within the log segments which have been truncated, we should not report an error. |
@nvartolomei fixed this in #18097 |
https://buildkite.com/redpanda/vtools/builds/13666
JIRA Link: CORE-2942
The text was updated successfully, but these errors were encountered: