[
https://issues.apache.org/jira/browse/IGNITE-8000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16450201#comment-16450201
]
Alexei Scherbakov commented on IGNITE-8000:
-------------------------------------------
Fixed as part of IGNITE-6827.
> Implicit transactions may not finish properly on unstable topology.
> -------------------------------------------------------------------
>
> Key: IGNITE-8000
> URL: https://issues.apache.org/jira/browse/IGNITE-8000
> Project: Ignite
> Issue Type: Bug
> Affects Versions: 2.4
> Reporter: Alexei Scherbakov
> Assignee: Alexei Scherbakov
> Priority: Critical
> Fix For: 2.5
>
>
> Add default tx timeout [1] to IgniteCacheMultiTxLockSelfTest test
> configuration.
> [1] c.getTransactionConfiguration().setDefaultTxTimeout(100000);
> Looks like in some case remote tx is added to rolled back version (because
> partition is gone) and subsequent near request for the same tx to this node
> fails.
> This is not happen if timeouts are disabled because corresponding check is
> skipped.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
