[
https://issues.apache.org/jira/browse/OAK-9656?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17469987#comment-17469987
]
Marcel Reutegger commented on OAK-9656:
---------------------------------------
PR https://github.com/apache/jackrabbit-oak/pull/453 adds a test that
reproduces the problem.
> Recovery runs mistakenly when system clock jumps ahead
> ------------------------------------------------------
>
> Key: OAK-9656
> URL: https://issues.apache.org/jira/browse/OAK-9656
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: documentmk
> Affects Versions: 1.6.0, 1.8.0, 1.22.0, 1.40.0
> Reporter: Marcel Reutegger
> Assignee: Marcel Reutegger
> Priority: Major
>
> Recovery may run mistakenly for other clusterIds when the system clock on one
> DocumentNodeStore jumps ahead. This is because from the DocumentNodeStore
> point of view/time the other clusterIds lease appear expired.
> The current implementation attempts to run recovery and fails because it
> eventually notices its own lease expired. It appears other clusterIds are
> modified anyway and then fail the next lease update because the leaseEnd does
> not match the expected value.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
