[ https://issues.apache.org/jira/browse/IGNITE-6827?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16425147#comment-16425147 ]
Alexei Scherbakov commented on IGNITE-6827: ------------------------------------------- [~agura], Thanks for comments. 1. There are no TODOs left by me in these classes. 2. This check is actually necessary here because we have special return value 0 if timeout is not set at all. 3. Fixed. I made an Upsource review for branch [1] Latest TC run [2] Some tests worth additional investigation. [1] https://reviews.ignite.apache.org/ignite/review/IGNT-CR-544 [2] https://ci.ignite.apache.org/viewLog.html?buildId=1176005&tab=buildResultsDiv&buildTypeId=IgniteTests24Java8_RunAll > Configurable rollback for long running transactions before partition exchange > ----------------------------------------------------------------------------- > > Key: IGNITE-6827 > URL: https://issues.apache.org/jira/browse/IGNITE-6827 > Project: Ignite > Issue Type: Improvement > Affects Versions: 2.0 > Reporter: Alexei Scherbakov > Assignee: Alexei Scherbakov > Priority: Major > Fix For: 2.5 > > > Currently long running / buggy user transactions force partition exchange > block on waiting for > org.apache.ignite.internal.processors.cache.GridCacheSharedContext#partitionReleaseFuture, > preventing all grid progress. > I suggest introducing new global flag in TransactionConfiguration, like > {{txRollbackTimeoutOnTopologyChange}} > which will rollback exchange blocking transaction after the timeout. > Still need to think what to do with other topology locking activities. -- This message was sent by Atlassian JIRA (v7.6.3#76005)