Save-while-active request does not work proper with application Servers

[Former Member]

Hello together,

I am using Save-while-active to save the Librarys R3SIDDATA and R3SIDJRN.

Everytime, not everytime but often, when we use two Windows application Servers together with the central Instance, the Save-while-active request can't find a "Save-while-active checkpoint". The backup end to the messsage "CPI8365 Save-while-active request requires commit or rollback for job 814325/QUSER/QXDARECVR" .

If I limit the usage of the application Servers to a few user, but let them run, the Save-while-active request will work.

The job 814325/QUSER/QXDARECVR holds a lot of locks. But I think this is normal, because this QXDARECVR Jobs are needed for applications Servers.

Is there a possibility to find out which locks a stopping the backup?

best regards,

Carsten Schulz

[christian_bartels]

Hi Carsten,

SAP Note 1061267 has suggestions how to analyze this kind of error. If there are any outstanding commit or rollback operations, you may be able to find them using the information in the note.

If you cannot find any outstanding commit operations for that job (at the time when the message was sent), there may be a bug in the save-while-active code of IBM. I noticed that you had already a problem with that in PMR 90245,070,724 (closed end of October last year). There the assumption was made that there were "orphaned" commit definitions, and the suggesion was to run:

- wrkcmtdfn job(*all)

- F11

- option 12 on all the commit definitions that had a value of YES in the Resync In Progress column

- are there any jobs related to these commit definitions?

Hope this helps. If you see this more often, you may have to open another PMR with IBM.

Kind regards,

Christian Bartels.

Edited by: Christian Bartels on Jan 20, 2010 5:14 AM

[Former Member]

Hello Christian,

yes, I know this Problem from the past.

We can't solve this problem last year.

There is no solution from IBM to this problem.

For that we decided last year not to use applications Servers any more.

Then we don't have further problems.

Now in 2010 we have new Kernel to the application Servers and we have now IBM i V6R1 with the newest patchlevel.

So I thought that we have a new chance to test this.

We tested it, an the result is, that we don't have a solution to that problem.

best regards,

Carsten Schulz

[christian_bartels]

Hi Carsten,

there may be no way around opening a PMR if this problem still exists. It is interesting to hear that this happened only with remote application servers, so maybe the problem is in the communication layer (called XDA).

SAP has provided a new commonication layer, which is known as "DBSL Direct Drive für DB2 für IBM i" and documented in SAP Note 1257635. Do you think you could try that without too much impact to your business?

Kind regards,

Christian Bartels.