on 11-25-2015 1:37 PM
Hi there,
I just noticed the following on transaction SM66:
The system is running fine. Looking in SM21 nothing seems to appear for the
same Process ID.
Is this something to worry about? How can I clear them down?
I went into SM50 and then restart process after error but it still stays highlighted in red.
Any ideas anyone?
Thanks
Andy
Hi Andy,
Please check the below notes, which gives you the various reason for workprocess restart
Note 101717 - Automatic restart of SAP R/3 work processes
And still if you are seeing the "red" highlights(In Failure column) in SM50. Please follow the below steps to resolve this:
tcode SM50>> List >> Reset >> Failures (After resetting, the red highlights will disappear)
Little late reply but still it may help someone who experience this issue in future
Best Regards,
Mervin Joseph
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Yes, You are right my friend, the above mentioned steps will only clear/reset the red highlights from SM50. As per my understanding I believe that is one of the query raised by Andy
Okay, lets dig little deeper now
Hi Andy,
After resetting the failure status if you are still getting the red highlight in SM50 then please share us few more info from the developer trace (dev_wx)file.
I know its not possible to attach the entire trace file since its big but what you can do is:
- Please make a note of the "failure reset time" (the time when you reset the failure list)
- download the trace file to your local desktop ( Select the wp with failure status from SM50 >> Administartion>> Trace >> Save as local file )
- Search for the word "error " which appears after the failure reset time. Please take a snap of the error logs and paste it here which will help us to investigate the root case for this issue.
- And also please check and share if you see any error log in tcode SM21 and ST22 at the same time.
Attaching a sample screenshot for your reference from dev_wx file:
Awaiting for your reply!
Best Regards,
Mervin Joseph
Hello Andy,
You could also reset the trace files (SM50 -> Administration -> Trace -> Reset -> work processes).
The next time you see a failure, open the corresponding trace file and search for " pid ".
Each failure means that the work process has restarted.
At each restart, the work process PID will change, and the work process will log the initial header at the trace again (that initial part that shows the system ID, kernel version, the PID, ...).
Once you locate the second occurrence of the PID (the first one would be the PID from before the new failure), copy the trace entries from around 100 lines up until around 100 lines down (from the line of the second PID) and attach it to this thread.
Notice that the failure that occurred at one work process might not be the same that occurred at all of them.
Regards,
Isaías
What function / ABAP report does the reset triggered from SM50?
I'd like to have a look at the source code
SM50>> List >> Reset >> Failures
(After resetting, the red highlights will disappear)
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Mervin,
Here we go, just checked SM66:
Now I went into SM21 and checked the log and seen that the WP's have since started
55 times:
When I now drill down for that WP: 21231
So I now went into the trace file and searched for the SP 21231:
Now I think I know what is causeing the error.
When you look at the first line it says 'ORACLE not available'
It just so happens that at this particulat time (and the repeating erors) is when our offline
backup is still running and so ORACLE will indeed be down.
Do you think this is the case? So I can just ignore them??
Many thanks
Andy
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Guys,
Yep it is definitely the offline backup causing these failures to appear in SM66. Just looked
now and they have started appearing again after the weekends backup.
I know I can ignore these but is there a job I can schedule to clear these down instead of having
to do this manually??
I like things nice and tidy !!
Many thanks for all your help it is much appreciated.
Andy
if possible restart the application when the system load is less .
hope this fix the issue .
Regards
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Andy,
If you are seeing no errors in the system log or WP logs, then you may find at the WPs are set to restart periodically - it's a profile parameter which sets this.
Regards,
Graham
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hey Graham,
I can see where you're coming from with your initial remark and probably by the screen shot with the number of failures across all WP being almost identical.
That said, the WP autorestart wouldn't (shouldn't) show up in the error count of a WP. You'd just see in the dev trace that the autorestart time had been reached and the WP would restart.
If you have a test system, check it out by setting it to a silly value like 60s.
Cheers,
Amerjit
Hi Andy,
The count tell the number of time the work process has restarted after the system restart.
In your case I see the main reason is heap memory. It is set low.
Work process restarts itself once it reaches heaplimit value or the heap configured for the work process
Parameters:
abap/heaplimit
abap/heap_area_dia
abap/heap_area_nondia
From your St02 screen-shot I can see the value set to heap is very small.
Regards,
Prithviraj.
Hi Andy,
Your SM66 screen shot is telling you that the <nn> WP have restarted <nn> times.
ie: The WP with PID 29906 has had 181 failures (restarts) but is currently in status WAITING (waiting for work).
What you need to check is the developer trace associated with the WP (SM50) and see from within the dev trace what errors are being logged.
Hope this helps you a wee bit.
KR,
Amerjit
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
86 | |
10 | |
10 | |
9 | |
6 | |
6 | |
6 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.