| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| WMB 6.1 initial deployment fails on 64-bit Linux |
« View previous topic :: View next topic » |
| Author |
Message
|
| Esa |
 Posted: Wed Jun 25, 2008 11:40 am Post subject: WMB 6.1 initial deployment fails on 64-bit Linux |
 |
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
Has anybody been successfully running Message Broker 6.1 on 64-bit Linux?
I have both configuration manager and broker running on same queue manager on 64-bit Centos 5.0 (Well, actually Centos is not supported, but it is supposed to be RHEL-compatible). I run WMB Toolkit on 32-bit Centos 5.1 on vmware on the same workstation. The workstation is Acer Ferrari 5005 with 64-bit AMD Turion dual core and 4 GB RAM.
I have struggled with this for months now. When I try to add the broker in the domain, the operation gets timed out. I have deleted both cfgmgr and broker at least 20 times (and dropped the databases, too). Now I have upgraded to 6.1.0.2 and I get more information in the system log. The execution group $SYS_mqsi creashes with a segmentation fault during the initial deployment. This happens several times with some seconds interval as the broker restarts the execution group until it thows the towel after about 5 tries.
The crach produces a core dump, a snap.*.trc and a javacore. So I think the JVM crashes.
I have run agent trace and component trace during the deployment but I cannot see anything useful there. Actually the component trace is empty, I guess the binary trace is still in a jvm-allocated buffer where the jvm crashes.
The last entry in the javacore is about com/ibm/broker/server/BrokerDatabase.getAllAclEntries (native method)
MEMINFO subcomponent contains:
Bytes of Heap Space free: e83f10
Bytes of Heap Space Allocated: 2000000
The free heap space value seem a bit susceptible. Could this happen because of heap overflow? The problem is that you only can increase jvmMaxHeapSize after the initial deployment, as far as I know.
Esa |
|
| Back to top |
|
 |
| smdavies99 |
| Posted: Wed Jun 25, 2008 10:48 pm Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
Have you checked to see if Broker has produces an FDC log of the event?
They should be somewhere around /var/mqsi/common/....
I had lots of problems with 64Bit V6.1 broker on both CentOS 5.1 & pukka RHEL 5.1. The FDC's reported problems allocating memory. Like you the server had plenty of RAM available and the kernel params were also correct (As per docs) and could run Oracle 11g without problems(not at the same time as broker I hasten to add)
In the end, we reverted to 32-bit V.6.0.0.7.
We still get an error every hour when the V6.1 ConfigMgr tries to get a 'report' from the broker. Still the broker is working.
This whole setup was for a POC so I didn't have the inclination to get involved with a PMR at this stage.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| Esa |
| Posted: Wed Jun 25, 2008 11:20 pm Post subject: |
|
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
Yes, I get abend files in the broker work path. Failed to allocate memory, as in your case.
And:
/opt/ibm/mqsi/6.1/lil/imbdfjps64.lil(Java_com_ibm_broker_server_BrokerDatabase_getAllAclEntries+0x138)[0x2aaab21532f8]
so it is not a jvm issue but the C library under JNI that fails to allocate memory.
I guess I will have to wait for the next fix pack. This is not a customer case so I think I cannot raise a PMR. Hope womebody else will.
Esa |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
