RGZPFM stopped working after OS upgrade to V5R4 from V5R3

[Former Member]

Hi Guys,

We recently upgraded the operating system on our DEV & TEST boxes from V5R3 to V5R4.

Before the upgrade, we used to use the program ZRGZPFM attached in SAP note number: 84081 Reorganization of database tables (RGZPFM) for Re-Organizing the database tables weekly after the offline backup.

Now When we submit this program with the parameters &LIBRARY= R3<SID>DATA & &THRESHOLD =1000, we get the following Error:

CALL PGM(*LIBL/ZRGZPFM) PARM(R3BWTDATA 1000)

Object MBRLIST in QTEMP type *FILE not found.

Output file MBRLIST created in library QTEMP.

Member MBRLIST added to output file MBRLIST in library QTEMP.

16108 records added to member MBRLIST in file MBRLIST in QTEMP.

String constant beginning "QDBRGZ_R" not delimited.

The fast copy, or initialization can not be done.

Function check. CPF3169 unmonitored by ZRGZPFM at statement 1800,

instruction X'004C'.

CPF3169 received by ZRGZPFM at 1800. (C D I R)

When we do a F1 above on the "String constant beginning "QDBRGZ_R" not delimited", it shows us the following message:

Message ID . . . . . . : SQL0010 Severity . . . . . . . : 35

Message type . . . . . : Diagnostic

Date sent . . . . . . : 02/11/07 Time sent . . . . . . : 18:31:49

Message . . . . : String constant beginning "QDBRGZ_R" not delimited.

Cause . . . . . : The string delimiter is missing in the constant beginning

with "QDBRGZ_R". The string is treated as if it were delimited by the end

of the source file.

Recovery . . . : Delimit the string constant. Check for any missing or

extra quotation marks and apostrophes. These errors are likely to cause

other errors. Some statements may not have been processed as the result of

either missing or extra string delimiters. Precompile the program again.

I have tried recompiling and generating the object ZRGZPFM after the OS upgrade...Also deleted the duplicate ZRGZPFM object from the library ZRPLOBJ

and checked after the re-generation that the object is the latest generated version...

I have checked the File QDBR0001 in library QRECOVERY....which is generated during the Re-org...Also tried deleting the file and then restarting the job but to no success....

The same program was running perfectly till Last week in V5R3...and the same program is running perfectly in our PROD box...for which we are planning to upgrade the OS this coming weekend..., which I guess we may have to defer...if this is not resolved....

just running out of ideas altogether...Cud it be a PTF or something else...??

Thanks for all your help guys...

Abhi

[christian_bartels]

Hi Abhi,

the message sounds a bit puzzling, because it is sent from SQ (-> SQL0010). The CL source does not contain SQL, and there is no reference to a string constant like "QDBRGZ_R" or similar. So my guess is that we either have a bug in the operating system, or some kind of mismatch in the program between V5R3 and V5R4. Can you try if the program works after recompiling it in V5R4?

If not, it would be interesting to see the call stack at the time of the error, to understand at what point the message is issued. If you submit your program to batch, you will usually see the job in MSGW when the error happens. At this point, you can answer the message (CPA0701) with option 'D' to get a dump of the CL vairables. This may help to understand the problem. You could also open a ticket at SAP (component BC-DB-DB4) to get this problem analyzed.

Kind regards,

Christian Bartels.

[Former Member]

Hi Christian,

Thanks a lot for the response. Yeah, even I was stumped as to what is "QDBRGZ_R"??

I think that this maybe QDBRGZ_R3<SID>DATA_/<file_name>, which it reads from probably library QRECOVERY/QDBR0001 at runtime. This QDBR0001 is created I guess,when the reorgs run.

and during the Re-Org, it is not able to read the proper file name(SQL error) and hence bombs out...My Theory....I do not know why although...???

not sure although...

Yes, tried, recompiling the program in V5R4 as I mentioned before in both my DEV/TEST boxes and the same result. Infact, both my boxes give exactly the same error, as mentioned before.

I did not guess a chance to Dump the CL variables as now the systems are up. But I will try tonight to do it. Will definitely post the dump.

Thanks

Abhi

[christian_bartels]

Hi Abhi,

can you please also try to get a complete joblog through DSPJOBLOG OUTPUT(*PRINT) or at least the information about sending and receiving programs (function key F9 on the message details)? It is possible that this is a bug in the IBM code, so it would be interesting to know who is sending and receiving the message here.

Kind regards,

Christian Bartels.

[Former Member]

Hi Christian,

Thanks again. I will do that tonight and will update you with the results...

I also want to test tonight if this thing happens only for a specific data library, which has a specific type of object or is system wide. I have noticed that it only gets stuck if the table name is of type "/BI00401" or "/BIC1408". I mean in general "/BI<number>" or /BIC<number>".

Maybe somehow its now able to parse out the table name when its trying to reorg the table. I will confirm it tonight, when i will run the Reorgs for a system which doesn't have these types of table in the data library.....

Thanks for all you help..

Abhi

[Former Member]

Hi Christian,

It finally turned out to be what I was suspecting. The Re-Orgs fail only, when the file name/table names have a "/" in them. The Re-orgs for my other systems go well as they do not have any table name starting with a "/".

So it looked like an IBM issue.

I did some search on the IBM support page and finally came across the cover letter for the PTF SI26293. This states that RGZPFM fails on file with '/' in filename. Hence I applied that PTF. After that ran the Re-Orgs and finished successfully.....

I wonder why SAP does not have this PTF on the latest infopar(II14126) for V5R4. It is a PTF without which reorgs cannot run in a BW system, where a lot of tables start with a "/".

It drove me crazy...actually.

Maybe if someone is reading this, can forward this to the infopar team.

Thanks a lot for all your help.

Abhi

[JanStallkamp]

Hi Abhi.

> I wonder why SAP does not have this PTF on the latest

> infopar(II14126) for V5R4. It is a PTF without which

> reorgs cannot run in a BW system, where a lot of

> tables start with a "/".

> It drove me crazy...actually.

> Maybe if someone is reading this, can forward this to

> the infopar team.

Thanks for pointing that out. We have given that information to the infoapar team and they will update it soon.

Best regards,

Jan