| Author |
Message
|
| NickB |
 Posted: Tue Jun 19, 2001 4:52 am Post subject: |
 |
|
Centurion
Joined: 20 May 2001
Posts: 107
Location: Zurich Financial Services
|
I'm having tremendous problems currently when deploying execution groups. I continually get error messages complaining about the execution group not responding in a timely manner. I've recreated the config qmgr to increase log sizes but this hasn't helped. I'ved checked the error logs on the AIX broker qmgr but again there is nothing forthcoming. syslog just repeats the BIP2066E error message and doesn't give me any more help.
I've stopped and restarted the broker and I've even gone as far as clearing down the SYSTEM.BROKER.* queues as the whole system got in a real mess.
Is there anything else I can look at? |
|
| Back to top |
|
 |
| zpat |
| Posted: Mon Jun 25, 2001 4:48 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
| Please post your MQSI level include CSD level, this applies to any problem report here really. |
|
| Back to top |
|
|
| craigd |
| Posted: Mon Jun 25, 2001 4:11 pm Post subject: |
|
|
Guest
|
| I had the same problem with an AIX broker. Often I could fix the problem by dumping out the MQSeries logs. |
|
| Back to top |
|
|
| shcherb |
| Posted: Wed Nov 14, 2001 7:37 am Post subject: |
|
|
Apprentice
Joined: 12 Sep 2001
Posts: 34
|
Hi,
I am having the same problem. What can I do ? How can I dump MQSeries logs ?
Please help ! It is very urgent.
Vladimir |
|
| Back to top |
|
|
| EddieA |
| Posted: Wed Nov 14, 2001 8:02 am Post subject: |
|
|
Jedi
Joined: 28 Jun 2001
Posts: 2453
Location: Los Angeles
|
Hi all,
I have seen a very similar problem at a customer I was working at on an AIX broker.
Basically, re-deploying message flows would throw the (one of) the execution groups into a CPU loop.
Next time this happens, run a 'top' or a 'ps' and see if any of the execution groups is sucking CPU. If it is, then you can 'kill' it and MQSI will recover. You may have to 'kill' all the execution groups for a full deploy. BEWARE. This does NOT clean up the database connections to the Broker database, so you may end up running out of handles.
If anyone is seeing this and it is repeatable, then PLEASE report it to IBM. My previous customer, for some reason, decided not to pursue this once I had finished my engagement with them.
Cheers,
_________________
Eddie Atherton
IBM Certified Solution Developer - WebSphere Message Broker V6.1
IBM Certified Solution Developer - WebSphere Message Broker V7.0 |
|
| Back to top |
|
|
| zpat |
| Posted: Wed Nov 14, 2001 2:15 pm Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
You may need to increase the max heap size:
Increase the max heap size for the Configuration Manager JVM to 256Mb:
Stop the Configuration Manager if it is running:
Via Services (Start -> Settings -> Control Panel -> Services) OR
Using the mqsistop configmgr command
Update the NT registry to set the JVM max heap size:
Start the NT registry (enter regedit at a command line)
Select HKEY_LOCAL_MACHINESOFTWAREIBMMQSeriesIntegrator2ConfigMgr
Add new string value MaxJVMHeapSize:
Edit -> New -> String Value
Rename to MaxJVMHeapSize
Right click and select Modify:
Enter 268435456 in Value data and click OK
Restart the Configuration Manager
Via Services (Start -> Settings -> Control Panel -> Services) OR
Using the mqsistart configmgr command
Restart the Brokers
Via Services (Start -> Settings -> Control Panel -> Services) OR
Using the mqsistartbroker command
-----
Make sure the MQ logs are big enough to store the deploy during the transaction. Add more secondary logs (since you can't change the size after install - thanks IBM!).
-----
If it's urgent - open a PMR with the support centre.
[ This Message was edited by: zpat on 2001-11-14 14:18 ] |
|
| Back to top |
|
|
| mpuetz |
| Posted: Mon Dec 17, 2001 4:16 am Post subject: |
|
|
Centurion
Joined: 05 Jul 2001
Posts: 149
Location: IBM/Central WebSphere Services
|
Hi folks,
the BIP2066 happened to us as well with 2.0.2 on AIX 4.3.3.
For us it was related to DB2 V7.1. The following steps needed to be done to fix it:
1.) increased dbheap to 1200
2.) increased app_ctl_heap to 1024
3.) most importantly upgrade to FP 4 !
(it is really a DB2 bug, the upgrade
is tricky, follow the instructions
with great caution. Should be done
in a testing environment first)
For large MQSI deploys you should eventually
monitor DB2 to set the parameters 1 and 2
correctly for your applications. The values
I quoted were sufficient for us, but they need not be sufficient for you.
Mathias Pütz
IBM-EMEA AIM Services
MQ/MQSI Specialist
|
|
| Back to top |
|
|
|
|
