[
https://issues.apache.org/jira/browse/OAK-786?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13703206#comment-13703206
]
Marcel Reutegger commented on OAK-786:
--------------------------------------
See also ConcurrentAddIT in oak-jcr for a test where SegmentMK is currently
ignored because of this issue.
> Fall back to pessimism
> ----------------------
>
> Key: OAK-786
> URL: https://issues.apache.org/jira/browse/OAK-786
> Project: Jackrabbit Oak
> Issue Type: Sub-task
> Components: core
> Reporter: Jukka Zitting
> Priority: Minor
>
> By default the SegmentMK uses optimistic locking when committing changes to a
> given journal. This works well most of the time, but there a few scenarios
> when this strategy will fail. For example:
> # A large batch operation while other smaller changes are being committed.
> # Lots of concurrent changes being committed against the same journal.
> In such cases optimistic locking will lead to some changes being delayed
> indefinitely, whereas pessimistic locking can guarantee eventual completion
> of all operations at the cost of potentially significantly reduced
> throughput. See http://markmail.org/message/dl3semmp577bnqoi for related
> discussion.
> As a hybrid solution that attempts to combine the best aspects of both
> optimistic and pessimistic locking, we should make the SegmentMK fall back to
> pessimism if optimistic locking fails after some N attempts or X seconds.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
