[ https://issues.apache.org/jira/browse/IGNITE-10520?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16748706#comment-16748706 ]
Andrew Mashenkov commented on IGNITE-10520: ------------------------------------------- There is no need to run test in ForceMvcc mode. This should be fixed. The original issue was that checkPutGet(ignite.cache(TX_CACHE)...) always failed if it had called right after checkPutGet(ignite.cache(MVCC_CACHE),...) Seems, it has already been fixed within some another ticket. > MVCC: Next to mvcc-tx non-mvcc-Tx failed with unexpected exception. > ------------------------------------------------------------------- > > Key: IGNITE-10520 > URL: https://issues.apache.org/jira/browse/IGNITE-10520 > Project: Ignite > Issue Type: Bug > Components: mvcc > Reporter: Andrew Mashenkov > Assignee: Ivan Pavlukhin > Priority: Critical > Labels: mvcc_stabilization_stage_1, transactions > Fix For: 2.8 > > > IgniteCachePrimarySyncTest.testPutGet) failed. > It run tx operations one by one > # tx on Transactional cache > # tx on Transactional_snapshot cache. > # tx on Transactional cache > Last tx failed with "TransactionException: Only pessimistic repeatable read > transactions are supported when MVCC is enabled." which is unexpected. > Looks like we forget to reset tx context somewhere. -- This message was sent by Atlassian JIRA (v7.6.3#76005)