[
https://issues.apache.org/jira/browse/ARTEMIS-3055?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17260221#comment-17260221
]
Justin Bertram commented on ARTEMIS-3055:
-----------------------------------------
The broker was shut down by what's called the "[critical
analyzer|http://activemq.apache.org/components/artemis/documentation/latest/critical-analysis.html].";
In short, the critical analyzer is designed to monitor how long particular
critical operations take to complete. If the any of the critical operations
take too long then the analyzer will take action (which is configurable). This
is the critical analyzer configuration from the default {{broker.xml}}:
{code:xml}
<!-- should the broker detect dead locks and other issues -->
<critical-analyzer>true</critical-analyzer>
<critical-analyzer-timeout>120000</critical-analyzer-timeout>
<critical-analyzer-check-period>60000</critical-analyzer-check-period>
<critical-analyzer-policy>HALT</critical-analyzer-policy>
{code}
This means that if any critical operation takes longer than 2 minutes (i.e.
120,000 milliseconds) then the broker will be halted (i.e. shutdown). The
critical analyzer is designed to prevent a "soft" outage, for example, when a
dead-lock prevents the broker from operating normally but all clients are still
connected and there are no obvious signs of a problem. If the broker is
properly configured as an operating system "service" then when the broker is
halted to clear the issue the OS will automatically restart it and clients can
quickly reconnect and resume normal operation.
According to your log, "path 0" in
{{org.apache.activemq.artemis.core.server.impl.QueueImpl}} took too long so the
broker was stopped. The queue in question was named
{{lumos-grouped-event-data-collection-app}}, and "path 0" is the operation
which adds a message reference to the tail of the queue. This operation is
performed, for example, when a client sends a message to a queue.
It's not clear why this operation would have taken long enough to trigger a
shutdown. Were there any other messages in the log which indicated any
potential problems? What was the state of the machine (e.g. CPU utilization,
heap utilization, etc.)? Can you provide your {{broker.xml}}?
> Server shutdown because the queue lock issue
> --------------------------------------------
>
> Key: ARTEMIS-3055
> URL: https://issues.apache.org/jira/browse/ARTEMIS-3055
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Components: Broker
> Affects Versions: 2.13.0
> Environment: artemis broker: 2.13
> client version: 2.10
> Reporter: Yuri Li
> Priority: Blocker
> Labels: features
> Attachments: artemis.log
>
>
> One standalone server is running in our production environment. after one
> week running, it shows the error as the log showing in the attachment.
>
> One problem may be that the client jar version is different the server
> version, I am not sure this is the root cause. just attach the log for your
> confirmation.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
