[ https://issues.apache.org/jira/browse/GEODE-1329?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15269679#comment-15269679 ]
ASF subversion and git services commented on GEODE-1329: -------------------------------------------------------- Commit b893abe094b2df73e51cdce1d0716fc984b1115c in incubator-geode's branch refs/heads/feature/GEODE-1255 from [~bschuchardt] [ https://git-wip-us.apache.org/repos/asf?p=incubator-geode.git;h=b893abe ] GEODE-1329 auto-reconnect attempts cease if kicked out during boot-up of the cache InternalDistributedSystem.reconnect() now includes cache-creation in its retry loop and, should the cache fail to start due to a CancelException it will shut down and try again. While creating a new test in ReconnectDUnitTest I found problems with the other tests in that class that I fixed. Notably the method getDistributedSystemProperties() wasn't returning the correct properties for a lot of test cases because the dsProperties variable wasn't being set by them and they were using the getCache() method. This was causing the current distributed system to be destroyed and a new one created with different properties than the test wanted. That was causing periodic test failures. > auto-reconnect attempts cease if kicked out during boot-up of the cache > ----------------------------------------------------------------------- > > Key: GEODE-1329 > URL: https://issues.apache.org/jira/browse/GEODE-1329 > Project: Geode > Issue Type: Bug > Components: membership > Reporter: Bruce Schuchardt > Assignee: Bruce Schuchardt > > If a server has been kicked out of the distributed system and is able to > reconnect it may cease efforts to rejoin the system if it is kicked out after > joining the system but before its cache is completely rebuilt. > The code in InternalDistributedSystem.reconnect() has a {{while}} loop that > exits after the new DistributedSystem has been created and then the method > goes on to rebuild the Cache. If there is a problem rebuilding the cache it > does not retry, but it should. -- This message was sent by Atlassian JIRA (v6.3.4#6332)