[infinispan-issues] [JBoss JIRA] (ISPN-1602) Single view change causes stale locks

[ https://issues.jboss.org/browse/ISPN-1602?page=com.atlassian.jira.plugin.... ] Dan Berindei commented on ISPN-1602: ------------------------------------ Erik, what's your JGroups configuration? I have seen an error after a merge, the new coordinator (dht10) was unable to get the list of running caches from dht11 (which started up as a separate partition). It seems that dht11 received the RECOVER_VIEWS command before the merged view was installed, and dht10 did not attempt to retransmit the message for over 1 minute (until the command timed out on the coordinator). Can you reduce {{STABLE.desired_avg_gossip}} in your JGroups configuration to 30000 and see if you still get the stale lock? We had a similar problem in our test environment but our retransmission delays were ~ 4 seconds so the merge never failed, it just took longer than usual. I started testing unicast messages for the RECOVER_VIEWS command to ensure that it arrives after the view installation messages, but I didn't reach any conclusion at the time. I have a hunch that the {{Message.OOB}} flag may also make it more likely for the message to be dropped, but I need to run more tests. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.jboss.org/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira