| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Core Dump When Processing 12 MB Message |
« View previous topic :: View next topic » |
| Author |
Message
|
| DJN |
 Posted: Fri Feb 11, 2005 5:32 pm Post subject: Core Dump When Processing 12 MB Message |
 |
|
Apprentice
Joined: 14 Apr 2003
Posts: 47
|
| We are still running version 2.1. We have a message flow which is processing a 12mb message. The flow has worked perfectly in test with smaller messages. Now with the larger message the flow is causing a core dump. The trace shows the last line of code executed in a SET using a LENGTH function. The function is trying to determine the length of the entire message. I have reviewed the code ( I didn't write it) and it looks fine. I don't believe the problem is in the code. My thoughts are that the broker or execution group is running out of memory but I don't see anywhere you allocate memory for a broker or execution group. I thought maybe the Java Heap may be the problem. Any ideas? |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Fri Feb 11, 2005 8:30 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Try upping the Java heap. There was a post earlier about removing a 10MB message from a queue.... Apparently the Java Heap did the trick for them....
Enjoy  |
|
| Back to top |
|
|
| shra_k |
| Posted: Sat Feb 12, 2005 2:49 am Post subject: |
|
|
Apprentice
Joined: 30 Dec 2002
Posts: 37
|
|
| Back to top |
|
|
| Craig B |
| Posted: Mon Feb 21, 2005 1:57 am Post subject: |
|
|
Partisan
Joined: 18 Jun 2003
Posts: 316
Location: UK
|
Hi,
If you have no Java nodes in your message flow, then increasing the Java heap size will not resolve the memory resource issue you are having. When processing large message bitstreams, it is not the bitstream itself that causes memory shortages, it is usually the size of the parsed message tree. It usually follows that the bigger the bitstream, the more message tree fields will be created. A good starting point is to monitor the size of your DataFlowEngine process to see what size it reaches before it abends. If you are running on AIX, then by default an Execution group has two segments giving it a maximum of 512MB. This number can be increased and can give a maximum address space size of 2GB. The maximum address space size on Windows, zOS and AIX is 2GB. On HP-UX it is 3.5GB, and on Solaris it is 4GB. Monitor your process size and if is abending before any of these sizes are reached then can you post more details of your configuration. If you are on AIX, then increase the number of segments your DFE has. Details of this can be found in the V5 documentation.
_________________
Regards
Craig |
|
| Back to top |
|
|
| DJN |
| Posted: Mon Feb 21, 2005 5:33 am Post subject: Thanks! |
|
|
Apprentice
Joined: 14 Apr 2003
Posts: 47
|
| Craig ... you were right on. We received a script from IBM last week that created 4 new DataFlowEngines. We found that the one sized at 1024 was able to process the message that failed. We are performing furthet tests to see the limitations of this engine. Thanks to all for your help. |
|
| 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
|
|
|
|
