[
https://issues.apache.org/jira/browse/IGNITE-27454?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18047756#comment-18047756
]
Roman Puchkovskiy commented on IGNITE-27454:
--------------------------------------------
Additional analysis showed that the locks are taken on the zone raft listener
level, so consistency is not compromised. We just need to cleanup the code
(remove additional locks on the table raft processor).
> Raft snapshot consistency is broken
> -----------------------------------
>
> Key: IGNITE-27454
> URL: https://issues.apache.org/jira/browse/IGNITE-27454
> Project: Ignite
> Issue Type: Bug
> Reporter: Roman Puchkovskiy
> Assignee: Roman Puchkovskiy
> Priority: Minor
> Labels: ignite-3
> Time Spent: 10m
> Remaining Estimate: 0h
>
> All commands in Raft listeners that MV storage must be executed under a lock
> acquired with PartitionDataStorage#acquirePartitionSnapshotsReadLock(). It
> seems that this was lost when moving to per-zone partitions as the only place
> where this lock is acquired is
> TablePartitionProcessor#onConfigurationCommitted().
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
