cancel
Showing results for 
Search instead for 
Did you mean: 

ORA-00020 after Kernel Update 700

denny-k
Participant
0 Kudos

Hi experts.

We are facing a problem with ORA-00020. We have systems running over years without such erros.

After setting a higher value, it takes a few days and we get the same error.


On Solaris 10 and Windows R2 we got ERP, BW and JAVA-Systems (700 up to 731..)


The problem results of a lot of zombie-processes. They are growing very fast.

On OS i cannot see who is creating them. Any ideas?

Accepted Solutions (1)

Accepted Solutions (1)

denny-k
Participant
0 Kudos

Thank you all for your help.

After a rollback to kernel 600 on solution manager the orphan-sessions are gone.

Thanks

Answers (4)

Answers (4)

Former Member
0 Kudos

Can you attach output of ps command with these defunct proccesses?

denny-k
Participant
0 Kudos

I wanted to do that but now, after stopping the solutionmnager, the zombie proceses are gone.

Our Solman is 7.1 and on a high release-level. That makes me wonder.

denny-k
Participant
0 Kudos

It figured out that the old kernel 500 has got the same problem. So the problem isnt the kernel.

We find out that the solution manager has no zombies - all other 16 systems do have.

ST-PI and co are uptodate. Any ideas?

No error queues in smq1 or smq2.

Brindavan_M
Contributor
0 Kudos

Hi,

Please run RTCCTOOL and update the ST-PI or ST-A/PI accordingly on your system.

Thanks,

Brindavan

denny-k
Participant
0 Kudos

Hi,

as i wrote: "ST-PI and co are uptodate".. RTCCTOOL is completly green. Thanks

divyanshu_srivastava3
Active Contributor
0 Kudos

Hi Dennis,

Can you check and reply on my 1st post and from Roman ?

Regards,

divyanshu_srivastava3
Active Contributor
0 Kudos

What you see on OS level are PIDs. Now you can map and find these PIDs on other levels, for eg, map this PID with work process PID to find what is the running process and so on. By doing this, we can get the application that is zombie in nature.

denny-k
Participant
0 Kudos

Our DBA and Unix-Admin told me that they are orphan-sessions. So they werent able to get infos about this sessions.

Now i rolled the kernel back to 600 and the problem is gone! thank you

Brindavan_M
Contributor
0 Kudos

This message was moderated.

denny-k
Participant
0 Kudos

Thanks, we had already increase the parameter. Didnt help.

Brindavan_M
Contributor
0 Kudos

Hi,

Are you getting any other issue like "ORA-01012: not logged on" as well? if your getting this issue then stop sap appliaction server and clear the shared memory also . Then try to login into oracle and excute the command  SQL> select status from v$instance;. if everthng looks fine then please start sap and then check.

Thanks,

Brindavan

denny-k
Participant
0 Kudos

No, we dont get Ora-01012.

divyanshu_srivastava3
Active Contributor
0 Kudos

What were the process after mapping the PIDs of these zombie processes ?

May be, your Solaris + kernel requires some patch work.

Regards,

denny-k
Participant
0 Kudos

After stopping the Solutionmanager the zombies are gone.

Our solaris and kernel patchlevel are fine.

What do you mean with your first sentence?