[
https://issues.apache.org/jira/browse/IGNITE-29000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18107280#comment-18107280
]
Sergey Chugunov commented on IGNITE-29000:
------------------------------------------
[~zstan], as far as I can see, callers of the method you modified don't check
its result for null assuming it is always not null.
So the patch looks safe from the standpoint of existing code and as it fixes
the test in question. From that perspective the patch is fine.
At the same time I don't fully understand what is so special about that test
and it may be a good idea to dig a bit further and understand WHY it reproduces
this behavior. Maybe we could find a deeper problem with SqlStatisticsCommand
itself along the way.
> NPE can be raised from
> GridCachePartitionExchangeManager#lastAffinityChangedTopologyVersion
> -------------------------------------------------------------------------------------------
>
> Key: IGNITE-29000
> URL: https://issues.apache.org/jira/browse/IGNITE-29000
> Project: Ignite
> Issue Type: Improvement
> Components: general
> Affects Versions: 2.18
> Reporter: Evgeny Stanilovsky
> Assignee: Evgeny Stanilovsky
> Priority: Major
> Labels: MakeTeamcityGreenAgain, ignite-2,, ise
> Attachments: image-2026-08-20-16-20-08-387.png
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> How to re-produce:
> run SqlStatisticsCommandTests from master branch, check commot test logs
> after (screen attached). This can mask further tests stagnation problems.
> !image-2026-08-20-16-20-08-387.png!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
