cancel
Showing results for 
Search instead for 
Did you mean: 

how to manually start a WP on OS level

Former Member
0 Kudos

Hi,

does anyone knows how to start a work process on OS level?

One of my WP "disappeared" and SM50 shows it as stopped and it cannot be restart from there...

Is there a way to restart it on OS level without restarting the whole SAP system?

thanks

Antonio

Accepted Solutions (0)

Answers (2)

Answers (2)

Former Member
0 Kudos

Hello Antonio,

I guess you should be able to restart the WP from SM50. Have you tried in the menu -> process-> cancel without core or restart after error?

Anyway, did you find the root cause of this?

Cheers,

Diego.

Former Member
0 Kudos

Hi,

yes, I tried all that...

On SM21 I get this error:

Operating system call fork:restart failed (error no. 22 )

I searched for the WP trace file and there's several "WP has reached abap/heaplimit" messages which forced SAP to restart the process.

My guess is SAP has reached the limit of restarts for this WP and a new restart is not allowed anymore.

I found this note which deals with similar problems (884359 - iSeries: as4/wp_max_restart parameter in instance profile) but I don't have the as4/wp_max_restart defined anywhere. RZ11 "tells me" the parameter is unknown...

Former Member
0 Kudos

Hello Antonio,

It doesn't matter that the parameter is unknow to rz11. Check for example [903245|https://service.sap.com/sap/support/notes/903245] (just a random note, not related to your issue)

What matters is that the parameter is known by the kernel. For example I added the parameter "rdisp/cleanup_after_crash" for another error and despite the rz11 warning it works.

Anyway, you'll need to restart SAP after the change, so this won't help you to solve the issue this time, it'll take effect after restart....

Cheers,

Diego.

Former Member
0 Kudos

Hi Antonio,

Coming back to your original question: There is no such way of starting a work process from the OS directly. As you might have noticed with a ps -ef, the dispatcher and work process share the very same code (you can see that they have the same process name), but the dispatcher is responsible for starting the work process and he has to flag internally that this new fork of this process will be an work process and not another instance of dispatcher.

About the heaplimit, this is not an error, it's just a cleanup task that the dispatcher does to clean heap memory usage. Everytime that your work process uses the value specified on abap/heaplimit, the work process will get restarted after it finishes it's activities.

If you are using an AS400, and your work process indeed restarts more than the as4/wp_max_restart, then the dispatcher indeed will not try to start this work process anymore. As the note states, if you are above the mentioned patch levels in the note, you can introduce this parameter on your profiles even tho it's not known in RZ11.

Cheers,

Maurício

Former Member
0 Kudos

Hi,

you can manage the workprocess using dpmon at OS level

syntax : dpmon pf= <Instance profile>

Regards,

Venkata S Pagolu