Long Running Dialog Processes

[Private_Member_19084]

Hi experts,

at christmas time we have done a migration.

Now we are on Linux (virtual) and ASE, EHP 6 and 7.31.

We are facing the problem now, that a lot of dialog processes gets hanging.

For example today I had again 3 on one instance, which are running since around 60k seconds.

How can I handle this issue or analyze it best?

I've found for example note 101717 which is describing the parameter rdisp/wp_auto_restart, but to be honest as 1st step I want to find out the root-cause.

The processes are not in PRIV....it seems like the are running, but the are doing nothing....
Hanging sometime also on a "Select-Request"

Kind regards

[alwina_enns]

Hello Christian,

do you have a screenshot about such processes in SM50? What do you see in the traces of those work processes? Please create a call stack for these work processes (SAP note 112), in the call stack you will see the last called funcitons, and if this call stack is always the same.

Regards
Alwina

[manumohandas82]

Hi Christian ,

Please confirm your Kernel release . Also check on the following Note

1885942 - Work process is assigned obsolete request

Thanks ,

Manu

[Former Member]

Hi Christian,

If you find any long running dialog work process under SAPSYST or SAPLSGUI report than yes it is a bug in kernel and you need to upgrade it.

Else better have a trace on the work process, runtiime trace SE30 can be handy in this situation. Also check if they are stuck on a particular semaphore (SEM column in SM50)

Regards,

Prithviraj

[former_member185954]

Hello,

What is the version of your kernel, I have faced a similar issue, however not sure if the solution is applicable to you ?

Regards,

Siddhesh

[Sriram2009]

Hi

1.  First think you need to find the which SAP program either standard or Z(Custom)  running long time?

2. Could you check the dialog process trace file is this any disconnection from the DB? any error message in SM21 or Dumps in ST22?

BR

SS

[Former Member]

plesae check the work process log for the process which is hung and if possible please attach here. Have you checked whehter the same program/tcode is causing the issue?

have you implemented rdisp/max_wprun_time parameter?

[Private_Member_19084]

Yes, the parameter is at around 900 seconds.

But I am sure, that the connection to the user is already disconnected.

I also have to kill these processes on os-level, and mostly with kill -9

I will check the log next time (maybe tomorrow) and also check on ASE-side, maybe I can see there something.

Kind regards