[ https://issues.apache.org/jira/browse/GEODE-4037?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16291681#comment-16291681 ]
Darrel Schneider commented on GEODE-4037: ----------------------------------------- I ran PersistentPartitionedRegionDUnitTest and do see it log the RETRYING_GII message. But it never looks like a hot loop. After the log message the next thing that does always does (it happened three times) is it then logs: Initialization failed for Region /__PR org.apache.geode.cache.persistence.PartitionOfflineException Also I don't see how this code would be in a hot loop because it nulls out "previousAdvice" before calling continue. So every other time around the loop it would skip this "if" and not log the message again. To proceed on this I think we will need to find a test that reproduces what Bruce saw. > hot loop in persistence advisor consuming CPU and producing lots of log output > ------------------------------------------------------------------------------ > > Key: GEODE-4037 > URL: https://issues.apache.org/jira/browse/GEODE-4037 > Project: Geode > Issue Type: Bug > Components: persistence > Reporter: Bruce Schuchardt > > While debugging a test failure I noticed that this code in > PersistenceAdvisorImpl was causing a hot retry loop that wrote a great deal > of output to the log file. > {code} > // Fix for 51698 - If there are online members that we previously > // failed to get a GII from, retry those members rather than wait > // for new persistent members to recover. > if (previousAdvice != null && > !previousAdvice.getReplicates().isEmpty()) { > logger.info( > > LocalizedMessage.create(LocalizedStrings.PersistenceAdvisorImpl_RETRYING_GII)); > previousAdvice = null; > continue; > } > {code} > I was going through a lot of tests at the time and didn't note which test hit > this but I _think_ it was PersistentPartitionedRegionDUnitTest. -- This message was sent by Atlassian JIRA (v6.4.14#64029)