[JBoss JIRA] Created: (JBESB-880) The ESB is blocked under heavy load when using JMS
by Jiri Pechanec (JIRA)
The ESB is blocked under heavy load when using JMS
--------------------------------------------------
Key: JBESB-880
URL: http://jira.jboss.com/jira/browse/JBESB-880
Project: JBoss ESB
Issue Type: Bug
Security Level: Public (Everyone can see)
Components: Rosetta, Transports
Affects Versions: 4.2 Milestone Release 3
Environment: RHEL 5, PostgreSQL
Reporter: Jiri Pechanec
Assigned To: Mark Little
Priority: Blocker
When the ESB is forced to process a huge batch of messages then it blocks completely.
The ESB configuration was just one service with one listener and gateway. Both gateway and listener are using JMS. The testing was executed twice using one and then four threads per listener and gateway.
The service contains two actions that serves only for message counting.
If you try to send 10000 messages sized around 20 KB it will probaly block after few minutes of processing. The ESB do not respond to TERM signal and even JMX console is not available. The only solution is to kill process violently. With the smaller messages the block does not happend or happends after bigger amount of messages.
The sender is generating following log messages
[Timer-1][BisocketServerInvoker] org.jboss.remoting.transport.bisocket.BisocketServerInvoker$ControlMonitorTimerTask@201d592a: detected failure on control connection Thread[control: Socket[addr=/10.34.34.47,port=2660,localport=51957],5,main]: requesting new control connection
We can not exclude problem in JBoss Messaging itself but when we tried to use plain JMS without ESB we were able to send/receive 300000 messages without any problem.
Moreover the more messages were processed the longer time it takes for the next batch to be processed.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
14 years, 4 months
[JBoss JIRA] Created: (JBESB-659) 2 JBossAS instances cannot share the same JBossMQ service for JMS gateways if JMS is installed
by James Williams (JIRA)
2 JBossAS instances cannot share the same JBossMQ service for JMS gateways if JMS is installed
----------------------------------------------------------------------------------------------
Key: JBESB-659
URL: http://jira.jboss.com/jira/browse/JBESB-659
Project: JBoss ESB
Issue Type: Bug
Security Level: Public (Everyone can see)
Environment: JBossAS 4.2GA, JBossESB - Head as of 07/05.2007.
Reporter: James Williams
Assigned To: Mark Little
I am trying to get 2 JBossAS instances to share the same JMS Server for gateways. There's a lot of moving parts, but I'll try to explain the configurations:
1. JBossAS 4.2.0-GA default config with all .ESB archives, minus the quickstart_helloworld.esb. This server is started using the "-b localhost -c node0" startup parameters.
2. JBossAS 4.2.0-GA default config with "jbossesb.sar" and "quickstart_helloworld.esb". All other .ESB archives are removed and my goal is to use the JMS service in server #1 for gateways/listeners. The quickstart_helloworld.esb archive in this server has its own gateway/listener queues. This server is started using the "-b 192.168.200.1 -c node1" startup parameters.
3. JBossAS 4.2-GA default config with "jbossesb.sar" and "quickstart_helloworld.esb". All other .ESB archives are removed and my goal is to use the JMS service in server #1 for gateways/listeners. The quickstart_helloworld.esb archive in this server has its own gateway/listener queues that are different from the gateways/listeners in #2. This server is started using the "-b 192.168.200.2 -c node2" startup parameters.
I'm not having any problems with nodes #2 and #3 sharing the JMS service that's running on #1, so long as JBossMQ isn't installed on node #2 or #3. If JMS is installed, quickstart_helloworld.esb archive fails to deploy on both #2 or #3. It complains that the Gateway queue does not exist, but it most certainly does. Worse yet, when I run the SendJMSMessage test class, it cannot find the queue either. It's as if jbossesb.sar is somehow screwing up all jndi lookups for the queue.
There are a couple things I've done to try to rule out other potential root causes:
1. I tried running #1, #2, #3 w/out any .ESB archives or jbossesb.sar to verify that a JMS client can indeed drop requests into queues on all servers. To me, this validates that 3 unclustered JBossAS instances can each have their own JMS queues that are all accessible on a single machine, like my test environment, just by having the app servers bootstrapped by unique config and ip address. I was worried that the JMS service might be the culprit, but that doesn't appear to be the case.
2. I removed JMS from #2 and #3 to verify that the queues are setup correctly on #1 and to verify that #2 and #3 can indeed share the same JMS service provided they do not have JMS running themselves.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
14 years, 8 months