[ https://issues.apache.org/jira/browse/NIFI-10974?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17680977#comment-17680977 ]
ASF subversion and git services commented on NIFI-10974: -------------------------------------------------------- Commit c3dd2569dbaddf11c1be114921e34cd9c96393a5 in nifi's branch refs/heads/main from Mark Bathori [ https://gitbox.apache.org/repos/asf?p=nifi.git;h=c3dd2569db ] NIFI-10974: Incorrect warning message on memory usage from MonitorMemory > Incorrect warning message on memory usage from MonitorMemory > ------------------------------------------------------------ > > Key: NIFI-10974 > URL: https://issues.apache.org/jira/browse/NIFI-10974 > Project: Apache NiFi > Issue Type: Bug > Reporter: Mark Bathori > Assignee: Mark Bathori > Priority: Major > Time Spent: 1h 40m > Remaining Estimate: 0h > > There are certain cases when the MonitorMemory reporting task is sending > incorrect warning messages on memory usage crossing the threshold. > {+}Repro steps{+}: > * In the MonitorMemory reporting task, set the threshold value to low enough > that the current memory usage is crossing that value. > * Restart NiFi while the monitoring task is still running. > * After the restart set the threshold to higher value above the actual > memory usage. > * The warning messages will be still send despite the threshold is above the > actual memory usage. > The main problem is that the *isCollectionUsageThresholdExceeded()* method > call on *MemoryPoolMXBean* have a *gcSensor* registered at the VM that's 'on' > attribute will get stuck on 'true' value. This will cause the method to > always return with exceeded. The sensor will get reseted on the next garbage > collection but until that it will send false messages. -- This message was sent by Atlassian Jira (v8.20.10#820010)