We recently upgraded to cognos 8.3. Last Friday our production environment crashed - it went into almost a zombie state where the content store was still communicating but users were getting a CAM-AAA-0071 error when trying to login. A recycle of the services resolved the issue but we crashed again this past Monday. The logs are filled the below error messages
xmlns:bus="http://developer.cognos.com/schemas/bibus/3/">
<severity>error</severity>
<errorCode>cmHeaderFault</errorCode>
<bus:message>
<messageString>CM-REQ-4159 Content Manager returned an error in the response header.</messageString>
</bus:message>
</bus:exception>
</detail>
170.6.137.4:9300 2316 2008-10-03 14:40:45.148 -5 Thread-20 caf 692 2 Audit.dispatcher.caf Request Warning received SOAP fault during capability check: details => <detail>
<bus:exception xmlns:bus="http://developer.cognos.com/schemas/bibus/3/">
<severity>error</severity>
<errorCode>cmHeaderFault</errorCode>
<bus:message>
<messageString>CM-REQ-4159 Content Manager returned an error in the response header.</messageString>
We pulled the box out that was the content manager and put in our 02 box. We are now starting to see the same errors again on the 02 box - not as frequent but we figure it took about 4 weeks for the 01 to start getting to that zombie state. Any insight would greatly be appreciated.
Can't say for sure, but looks like the memory leak issue - check with Support - some things I've heard people do: use cognos.cgi instead of cognosisapi.dll; smaller memory model; less memory allocated in the memory model; periodic reboots or restart of services
Hi,
I see the same error in my logs after I upgraded to 8.3. We noticed thousands of CM-REQ-4159 errors in the logs. They seem to start out in low numbers but start to increase and become more frequent until we finally see the CAM-AAA-0071 error and no one can log in. I did the same thing and pulled 01 (primary CM) out of rotation last Wed the 8th and the system is stable since - I did see the CM-REQ-4159 errors appear again, but not nearly as frequently. Do you have any updates on this issue? Any help would be much appreciated. I have a high priority ticket open with Cognos, but they have not provided any insight as to what could be happening....
Thanks!
Someclown - thanks for your response - is this memory leak for IIS - we are using websphere - any insight would be awesome
I have nothing definitive, but I think and I've heard that it's java based - made worse with cognosisapi.dll (for Windows users). So, my guess is that you still are eligible for the issue with Websphere
dstruck -
Do you still have issues?
If yes please read on....
Content Manager Jave.exe is not releasing memory. This will be more worest as you have more concurrent usage.The components that making JNI calls to LDAP is causing this. There is a patch with cognos support which will resolve, rather I say delay the memory leak. There are some other settings like HeapDecommitFreeBlockThreshold needs to be done along with patch. More details on HeapDecommitFreeBlockThreshold at http://support.microsoft.com/kb/315407
Let me know if you need more details.