cancel
Showing results for 
Search instead for 
Did you mean: 

Remove failed archiving process of business process instances

Former Member
0 Kudos

Hello,

I started an archiving process (bpm_proc_write) by selecting process instances in 'manage processes' --> archive. This archiving process failed and since then gets rerun every week and fails again each time. Why is it rerun? I didn't configure to rerun this archiving process. I deleted all entries for this archiving run in 'archiving monitor' and 'java scheduler jobs'. However this archiving run was rerun and caused again server crash. How can I stop/remove/get rid of this archiving job?

Thanks for hints.

Anja

Accepted Solutions (0)

Answers (2)

Answers (2)

0 Kudos

Hi Anja,

Did you got a response on the above question concerning "how to change the parameters of the bpm_proc_write job"? I have the same situation as you describe and I would like to solve it.

Kind Regards,

Marcel Knijnenburg

Former Member
0 Kudos

Hi Marcel,

unfortunately not. We stoped bpm archiving as it caused too many problems (OOM due to very large process instances, but selecting manually the propper amount and size of instances is not realistic; running archiving jobs seem to cause to fail the bpm timer event: we had huge amount of instances whose timer event didn't expire any more and whose next expiration time should have happened during the time range when the archive job actually was running).

What is your experience with bpm archiving? Do you have a few hundred process instances to archive on a daily basis and is this done fast and reliable?

Regards

Anja

0 Kudos

Hello Anja,

I solved the issue by following the hints in http://scn.sap/com/message/16811258. We daily archive appr. 3000 completed process instances and that takes about 3 seconds (running HANA).

Kind Regards,

Marcel

Former Member
0 Kudos

Hi Anje,

What is system version and SP? There are many improvements in later SPs of both 7.40 and 7.31. A common problem I've seen is that the parallel archiving configuration can be problematic in a small cluster.  OOM during archiving is typically due to too many jobs running on a single Java node or when too many process instances are put into a single job.  Both of these situations are avoidable with the right configuration.

regards, Nick

Former Member
0 Kudos

Hello Anja,

Can you provide any hints about the job failure from the job scheduler logs themselves? Also check to see if any Task remains for this job. What is your release/version?

regards, Nick

Former Member
0 Kudos

Hi Nick,

I already removed the entries from the job scheduler tab 'jobs'. There were/are no entries in tab 'tasks'. We're running NW 7.31SP06 (BPEM-CORE SP06PAT006).

I just noticed that some other jobs (successful archive jobs) were rerun, too. Some were rerun (once, twice or three times), others were not. I cannot see a pattern, why some did and others not (I cannot tell whether the rerun jobs where scheduled archive processes). May it somehow be related to the CleanJobTask_Default job? May it be that an archive job doesn't get properly cleaned up and the  CleanJobTask_Default doesn't detete this archive job as its retention time hasn't yet expired? But why are some of these archive jobs rerun only once? How can I change the retention period for bpm_proc_write?

Thanks.

Regards

Anja