[
https://issues.apache.org/activemq/browse/AMQ-2557?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=56762#action_56762
]
Gary Tully commented on AMQ-2557:
---------------------------------
rob just merged that fix to the 5.3 branch r896063 so it will be in tonight's
snapshot.
> Message Broker Hangs When Temp Store Fills
> ------------------------------------------
>
> Key: AMQ-2557
> URL: https://issues.apache.org/activemq/browse/AMQ-2557
> Project: ActiveMQ
> Issue Type: Bug
> Components: Broker
> Affects Versions: 5.3.0
> Environment: AMQ 5.3, Open Suse 11.0, JVM 1.6.0_14-b08, KahaDB,
> default store cursor, temp storage set to 300mb.
> Reporter: Joe Fernandez
> Fix For: 5.4.0
>
> Attachments: amq.out
>
>
> This is an exception test with the following set up.
> ProducerTool sends 200000 non-persistent messages to a Topic with the message
> size = 2000.
> The ConsumerTool that reads the messages from the Topic has a sleep time of 2
> msec, thus creating a slow consumer scenario.
> Temp store = 300mb, Topic memoryLimit = 30mb, producerFlowControl = false.
> The consumer begins to receive messages and messages get written out to the
> temp store as expected, because the slow consumer cannot keep up with the
> message load and temp topic fills up. However, when the temp store eventually
> fills as well, the broker goes comatose. It no longer accepts clients
> connections and stops dispatching messages to the slow consumer. You'd
> expect the broker to at least continue dispatching messages to the slow
> consumer while the producer is blocked awaiting the temp store to free up. If
> I double the temp store capacity, everyone is a happy camper. I'll attach a
> thread dump.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
