Management Kernel not coming up after TS upgrade to 8.98.2.0

CNC Guy

Well Known Member
Folks,

We are in the process of Ugrading from Tools 8.96 D1 to 8.96.2.0. We have already upgraded the test environment i.e. DV,PY,UA and currently testing before we upgrade the PROD toolset.

After the TR upgrade when we start JDE services on the Enterprise Server via the Server Manager 2 kernels go to zombie automatically after start i.e. die just after starting namely: MANAGEMENT KERNEL and METADATA KERNEL.

Since the MANAGEMENT SERVER KERNEL is not getting up when we restart the JDE services we are unable to see any Runtime metricis for the Enterprise Server via SM console.

Surprisingly though we can start and stop jde services of that enterprise server through the SM which makes me think the agent running on that server is fine.

Any clues? we have even logged a case with oracle but as you know they take all the time..

we even tried starting from the unix itself (as opposed to doing it from SM) but result is the same. The debug file for the kernel is attached.

Thanks,
CNC Guy
 

Attachments

  • 153188-jdedebug Management Kernel.txt
    10.1 KB · Views: 482
[ QUOTE ]
Folks,

We are in the process of Ugrading from Tools 8.96 D1 to 8.96.2.0. We have already upgraded the test environment i.e. DV,PY,UA and currently testing before we upgrade the PROD toolset.

After the TR upgrade when we start JDE services on the Enterprise Server via the Server Manager 2 kernels go to zombie automatically after start i.e. die just after starting namely: MANAGEMENT KERNEL and METADATA KERNEL.

Since the MANAGEMENT SERVER KERNEL is not getting up when we restart the JDE services we are unable to see any Runtime metricis for the Enterprise Server via SM console.

Surprisingly though we can start and stop jde services of that enterprise server through the SM which makes me think the agent running on that server is fine.

Any clues? we have even logged a case with oracle but as you know they take all the time..

we even tried starting from the unix itself (as opposed to doing it from SM) but result is the same. The debug file for the kernel is attached.

Thanks,
CNC Guy

[/ QUOTE ]

If I recall correctly those two kernels use JVM so it makes sense that both are failing. Something is not right with Java - classpath or something like that, perhaps the JVM settings in the jde.ini file.
 
Thanks brother_of_karamazov that is what we feel as well looking at the logs and yes these 2 kernels do use JVM. But we are not sure what are we missing.

I am attaching our jde.ini from server as well. Can you help?

One thing though : we have installed the server manager agent on the unix enteprise server on a differnt folder (a differnt mount point) that our active jde directory.

Thanks,
Shashank
 

Attachments

  • 153241-testjde.txt
    13.6 KB · Views: 186
Hi,

You add in the entries following the solution from Oracle Support as below:

SAR #8735811 fixes this issue for Server Manager. The fix was introduced with version 8.97.1.3 of the Server Manager.


WORKAROUND:
Change the JDE.INI file via the Server Manager Console so that it has the correct settings:

[JDENET_KERNEL_DEF31]
dispatchDLLName=xmlpublisher.dll
dispatchDLLFunction=_XMLPDispatch@28

[JDENET_KERNEL_DEF32]
dispatchDLLName=jdemgmt.dll
dispatchDLLFunction=_ManagementDispatch@28
 
[ QUOTE ]
Folks,

We are in the process of Ugrading from Tools 8.96 D1 to 8.96.2.0. We have already upgraded the test environment i.e. DV,PY,UA and currently testing before we upgrade the PROD toolset.

After the TR upgrade when we start JDE services on the Enterprise Server via the Server Manager 2 kernels go to zombie automatically after start i.e. die just after starting namely: MANAGEMENT KERNEL and METADATA KERNEL.

Since the MANAGEMENT SERVER KERNEL is not getting up when we restart the JDE services we are unable to see any Runtime metricis for the Enterprise Server via SM console.

Surprisingly though we can start and stop jde services of that enterprise server through the SM which makes me think the agent running on that server is fine.

Any clues? we have even logged a case with oracle but as you know they take all the time..

we even tried starting from the unix itself (as opposed to doing it from SM) but result is the same. The debug file for the kernel is attached.

Thanks,
CNC Guy

[/ QUOTE ]
What's the Logs show on the zombie kernel? I have to assume the logs whould give some error. The JVM point is a good one. You can point the JAVA_HOME at the jave instance in JDE or install a JDK.
 
Back
Top