When network activity is high, the worker thread can turn into turbo mode or read several requests in a single read. If the request is a persistent search, refcnt should not be decrease if we are in turbo or more_data mode. Indeed in such case the worker thread keeps accessing the connection and when it will end turbo or more_data it will decrease again the refcnt.
So the connection can be errounsly removed from the active list
All versions
testcase will be provided
It can create messages like
ERR - connection_release_nolock_ext - conn=6068 fd=66 Attempt to release connection that is not acquired
And in some rare case lead to a crash
No warning or crash
Metadata Update from @mreynolds: - Custom field component adjusted to None - Custom field origin adjusted to None - Custom field reviewstatus adjusted to ack - Custom field type adjusted to None - Custom field version adjusted to None
8f04487 -> master a2e752b..42ab464 389-ds-base-1.3.8 -> 389-ds-base-1.3.8 f15396a..2ce5261 389-ds-base-1.3.7 -> 389-ds-base-1.3.7
Metadata Update from @tbordaz: - Custom field reviewstatus adjusted to None (was: ack)
Metadata Update from @tbordaz: - Issue close_status updated to: fixed - Issue status updated to: Closed (was: Open)
389-ds-base is moving from Pagure to Github. This means that new issues and pull requests will be accepted only in 389-ds-base's github repository.
This issue has been cloned to Github and is available here: - https://github.com/389ds/389-ds-base/issues/2827
If you want to receive further updates on the issue, please navigate to the github issue and click on subscribe button.
subscribe
Thank you for understanding. We apologize for all inconvenience.
Metadata Update from @spichugi: - Issue close_status updated to: wontfix (was: fixed)
Log in to comment on this ticket.