Scheduled Jobs execution issue

[Nikhil_Makhija]

Hi All,

I scheduled few jobs though these jobs are running successfully and I can see their updated next run time but in real they are not executing what I mean here is that jobs are not updating data to tables or NW Queue or some logic to write to ECC etc.

If the same BLS mapped in job executes manually from workbench it updates all the tables/queues and run the logic successfully but its not executing the logic written in BLS automatically through jobs.

Is there any pointers why its happening ? Here are some details for reference-

SAP MII 12.2.4 Build(92)

OS : Linux

DB : Oracle 11g

NW : 7.3

Thanks

Nikhil

[Former Member]

We faced the issue.

We changed the Persist Transaction: Always and it worked. This issue was solved with recent SP.

This setting is not recomended in Production system

[Former Member]

Hi @Pradip..

If this setting is not recommended in Prod.. Is thr any other way out. ??

@All...

Not only scheduler.. Has anyone faced a below scenario ..

Trx A has web service ..and  Trx B is calling trx A.

In Workbench..

1. Run TrxA ..it works fine..

2. Run TrxB shows execution of Trx A positive (seen in tracer) but actually it fails to call web service.. ( algh web service returns .success =1 )

@All.. If its a know issue.. is there any note for same..

Thanks All.

Ruchir

[Nikhil_Makhija]

Hi All,

Thanks for your comments , well now the issue seems to be resolved and jobs are executing fine from system automatically there was some update happened on the server from Basis team not sure if it was security or patch update which corrected this issue for now.Will post more  details once I get the rootcause.

Ruchir,

It happens sometime that from MII connection has been established successfully to webservice but webservice in turn failed to connect with its internal database so in that case you will get success as 1 but in response you won't get  XML structure or incomplete XML structure.

Please post more details if this case  is not applicable for you with details like response you get or  error message you get in response

Regards

Nikhil

[Former Member]

Hi Nikhil,

I faced the same problem for the HTTP post from MII. I saw that when its executing from scheduler its not working but when I run it manually from workbench its working fine.

The resolution for my issue was the User we have used with the credential alias was expired, so when it was running from the scheduler it was unable to access, but when we executed it from the workbench then it fetched the credential from the system session and executed perfectly. So when we changed the user to active then it worked perfectly from scheduler.

So I am not sure that your issue is similar or not, but you can check that point once also.

Regards,

Suman

[Former Member]

@Hari.. to see if job is executing fine.. as Nikhil said..job works sometime and sometime doesnt.. we can check our db table to see if expected data from job updated the table or not..

I hope that should also work? Or if something more can be extracted from se37.. pls guide on that as to wht more to look for....

@Nikhil .. try to see the transaction manager or check nw logs to see if you can find something useful if particular scheduled run failed in btwn for some reason..

[Nikhil_Makhija]

Thanks for your response .

@Mike : Most Jobs are calling BAPI/RFC and updating data in to relational tables and some are calling webservices to update data in MII.

Yes the jobs are executing fine manually through scheduler. Automatic execution of jobs are successful sometime and sometime not.

URL seems to be pointing fine .Also Job history seems to be showing execution time for last 3 runs only don't know why  and taking very less time as compared to manual execution.

I am checking for any pattern or pointers,will update if able to find rootcause.

@Hari : Yes I am debugging and will try your suggestion as well.

Thanks again for your help.

Regards

Nikhil

[agentry_src]

Execution time is usually much less from Scheduler than in the workshop since there is a fair bit of overhead in loading and maintaining the transaction in the Workbench.  Not to mention the additional logging...

Good luck and please post your result when you find the solution.  I think Hari probably hit one of the most likely reasons for the problems.

I am a little confused by your statement about calling webservices to update MII.  Can you provide more detail?

Regards, Mike

[HariCS23]

Hi Nikhil, One thing you can try if the job is in development system is to put external break  point in RFC (SE37) to see job is executing properly.

Note: User in Credentials store should be same as user who logged in to ECC to start the external debugging.

Thanks

Hari

[agentry_src]

Hi Nikhil,

Can you be a bit more specific on what the jobs are doing? 

Do the same jobs execute properly when manually executed through Scheduler?

Do you have any urls which are locally defined which might be working when working locally (workbench) and pointing to the wrong places when executed from the server?

Just basic questions, but maybe they will point you in the right direction.  More details of what you are doing may also help.

Regards, Mike