Reclaim Storage needed after kernel update

[Former Member]

Last weekend we moved from a 620 to a 640 kernel for our PRD instance. Our TST instance, on the same system was moved weeks ago.

The PRD ran without problems after the upgrade to the new kernel.

In the backup we do a Save While Active of SAP data, except on Wednesday. So before last night's backup SAP was stopped and after the save started again. In the history log we have several messages with msgid CPF32A6 "Cross-reference notification 1190 for request 54, problem type 1." Which requires a RCLSTG to be run.

Does someone have an explanation for this behavior ?

[Former Member]

Hi R.,

it has nothing to do with your kernel change !

From time to time, the X-Ref files are running out of sync . this is unfortunately not predictable and reproducable. When the system detects such a mistake, as it did on your site, this message is created. You should now run a fast RCLSTG as follows when SAP is down:

RCLSTG *DBXREF

If you would use *ALL - the default - it might take 8-32h. This version just takes 30 - 60 minutes.

Regards

Volker Gueldenpfennig, consolut.gmbh

http://www.consolut.de - http://www.4soi.de - http://www.easymarketplace.de

[Former Member]

Hello Volker,

Sorry for leaving my name out, it is Rudi by the way.

After running the RCLSTG *DBXREF, I started SAP again. In the history log of the system the same message popped up again. This messages indicates a new RCLSTG is needed.

This very same problem was experienced by me before. An update of the DBSL in the Kernel library solved this issue. At that time the symptoms were the same, after a RCLSTG was run SAP was started producing a message indicating a new RCLSTG was needed.

I have been working with the i5/iSeries/AS400 since the beginning back in 1988, but I have never came across systems asking for a RCLSTG so many times as SAP does. I have the feeling there is more to it then just a coincidence.

What I will do is open a PMR for this issue and update this forum with the outcome of it.

Kind regards, Rudi.

[christian_bartels]

Opening a PMR with IBM is probably the right thing to do in this case; however, it may be difficult for IBM to find the root cause of this error.

If you look at the details of the CPF32A6 message, you can see details like "Object Context : R3<sid>02060", "Object Name : ZPERFTESHV", "Cross-reference File : QADBPKG", and "Request Job Name : SAPSTRSRV" (V5R4 only). The actual values in your case will be different, I took these examples from one of our systems. These values (object context and QADBPKG) indicate that the problems happened with SQL packages, and on my system it seems to me that two jobs where trying to delete the same SQL packages at the same time. This did not cause any error other than the message in QHST, so it does not seem to be necessary to run RCLSTG. In your case, if the message details are different, or if you see "real" errors (short dumps or messages in the developer trace), the situation may be different though. If you have the same symptoms as we do, you may want to check SAP note 793279.

Kind regards,

Christian Bartels.