[
https://issues.apache.org/jira/browse/IGNITE-28758?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18088260#comment-18088260
]
Nikita Amelchev edited comment on IGNITE-28758 at 6/11/26 3:30 PM:
-------------------------------------------------------------------
Merged into the master.
[~oleg_vlsk], Thank you for the initial contribution.
[~alexpl], Thank you for the review!
was (Author: nsamelchev):
Merged into the master.
[~alexpl], Thank you for the review!
> Explicit lock local-DHT callback can race with timeout/cancel cleanup.
> ----------------------------------------------------------------------
>
> Key: IGNITE-28758
> URL: https://issues.apache.org/jira/browse/IGNITE-28758
> Project: Ignite
> Issue Type: Bug
> Reporter: Nikita Amelchev
> Assignee: Nikita Amelchev
> Priority: Major
> Labels: ise
> Fix For: 2.19
>
> Time Spent: 40m
> Remaining Estimate: 0h
>
> For non-transactional explicit locks, the local DHT lock callback may
> complete after the outer GridDhtColocatedLockFuture has already been finished
> by timeout/cancel/error cleanup. In that case cleanup removes the explicit
> lock candidate, but the late local callback still proceeds to
> markExplicitOwner(), which can hit GridCacheExplicitLockSpan.markOwned()
> assertion or leave a locally acquired lock behind.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
