[
https://issues.apache.org/activemq/browse/AMQ-2564?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gary Tully resolved AMQ-2564.
-----------------------------
Assignee: Gary Tully
Fix Version/s: 5.4.0
Resolution: Fixed
r 946974 should resolve this. One thing to note is to ensure that the usage
limits on the slave are a little higher than the master such that flow control
happens only on the master, other wise replication can block which can block
message consumption.
the revision provides a test and resolves a hang situation where the producer
missed a reply because the replication messed with the correlation id. Normally
the ids are in sync, but when a message is queued waiting for space, they can
be mismatched. having the replicator (MasterBroker filter) copy the underlying
message resolved this issue,
> Pure Master Slave broker hangs when storage limit is reached
> ------------------------------------------------------------
>
> Key: AMQ-2564
> URL: https://issues.apache.org/activemq/browse/AMQ-2564
> Project: ActiveMQ
> Issue Type: Bug
> Components: Broker
> Affects Versions: 5.3.0
> Environment: Ubuntu 9.10 Desktop, Java 1.6 (build 1.6.0_16-b01)
> Reporter: Jörn Kottmann
> Assignee: Gary Tully
> Priority: Minor
> Fix For: 5.4.0
>
> Attachments: activemq-master.xml, activemq-slave.xml
>
>
> The pure master slave broker hangs when the storage limit is reached and
> dequeueing messages from the queue is not possible.
> The behavior can be reproduced with a producer which sends messages to the
> master broker until the slave broker writes
> "INFO | Usage Manager Store is Full. Stopping producer
> (ID:karkand-51292-1263305705067-0:0:1:1) to prevent flooding
> queue://TestQueue. See http://activemq.apache.org/producer-flow-control.html
> for more info"
> this message to the console.
> To reproduce the store is full message on the slave broker maybe 2 - 3
> attempts are necessary otherwise
> the messages appears on the console of the master broker.
> Any attempts by a test consumer to dequeue messages fail, it just hangs.
> The following information gather with JMX might be interesting:
> MASTER: StorePercentUsage=99
> SLAVE: StorePercentUsage=100
> The attributes for the Queue are identical on both
> brokers.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
