The delete job SAP_BC_XMB_DELETE_<CLNT> takes long time to move messages

[jegathees_waran]

Hello Friends,

In our production system, there is one copy job SAP_BC_XMB_DELETE_<CLNT> which runs everyday to switch the container.  Previously it was taking on an average 3 hrs to transfer (copy) 6 Lakh messages from one container to another (SXMSPMAST to SXMSPMAST2 and Vice Versa on the next run).  It transferred on an average 15000 messages for every 5 min and complete the job.

Since last sunday, this job transfers 2200 messages only in every 5 min. Due to this,  it is taking lot of time to complete the job. (to transfer 6 or 7 Lakh message). It took 30 hrs to complete the job yesterday.

We are not able to find the reason, why suddenly the no. of message transfer is decreased. On what basis it reduces the message count.  We checked our table space, DB update statistics. Everything is fine. No messages are stuck in SMQ2. Message Processing in SXMB_MONI is fine.

In Production system, Oracle 11g DB is used.

Please clarify what could be the problem?

Kind regards,

Jegathees P.

[baskar_gopalakrishnan2]

Since this is happening recently and you already made sure oracle disk space is not the issue at all, I believe any other process or batch job running during that time slows down this. Please monitor the performance of the server including number of threads running and processor memory and so.. This might give you some clue about slowness.  If you are archiving in file system, check whether file disk is not the issue.

[jegathees_waran]

Hi Baskar,

Recently we added one Java Server Node in Production System and our basis team done some Oracle Parameter changes. Will this affect SAP ABAP tables or impact the performance of copy job. Other background jobs are running fine. If you have any clue, kindly tell me.

[baskar_gopalakrishnan2]

Without seeing the parameter changes in oracle, it's very hard to confirm. But certainly things that were working and suddenly a big change in performance then we really need to consider the very latest changes or addition in terms of servers, hardware etc.   You might want to ask Basis or DBA to compare the performance of current log files with past weeks and see whether recent parameter changes causing this issue.

[iaki_vila]

Hi,

Have you checked the note 872388?:

If you encounter performance problems after execution of the switch this  might be due to deletion of the table metadata during the context switch. A solution for this problem is described in Note 1032733.

Then, read carefully the note 1032733

Let us to know if these notes helps.

Regards

[jegathees_waran]

Hi

We have already gone through the note 1032733.  The correction delivered in this Note for SAP_BASIS version 640, 700. Our SAP_BASIS release is 711. So, this will not be applicable for us. Also the Sub Notes which are mentioned for Oracle in this Note, not related to this problem.

So, we doubt about the index quality of the tables.