[ https://issues.apache.org/jira/browse/IGNITE-10434?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16729454#comment-16729454 ]
Igor Seliverstov commented on IGNITE-10434: ------------------------------------------- [~vozerov], The answer on both questions - currently lazy=false forces for such DML requests, initially there were dozens of problems with H2 cursors and connections, so, we weren't able to detach and reattach query processes from one to another thread gracefully. As first implementation lazy=false just was forced. this means: 1) only first call is time consuming, which executes the query and prepares full set of rows, each subsequent call is non blocking, so, it executes in the cache pool. 2) as far as there is no lazy execution there is no lock related problems at now. A fair lazy execution implementation is addressed for further releases. IGNITE-10829 is raised for this activity. > MVCC: Transaction asynchronous rollback bug. > -------------------------------------------- > > Key: IGNITE-10434 > URL: https://issues.apache.org/jira/browse/IGNITE-10434 > Project: Ignite > Issue Type: Bug > Components: mvcc > Reporter: Roman Kondakov > Assignee: Igor Seliverstov > Priority: Major > Labels: mvcc_stabilization_stage_1, transactions > Fix For: 2.8 > > > There is some bug in mvcc tx asynchronous rollback flow. Sometimes > transaction is not rolled back completely: Remote transactions remain alive > while Near and Dht transactions are terminated locally and on MVCC > coordinator. > Reproducer: {{MvccTxRollbackAsyncTest#testMixedAsyncRollbackTypes}} -- This message was sent by Atlassian JIRA (v7.6.3#76005)