IBM Mainframe Forum

by **NicC** » Thu Aug 11, 2011 5:22 pm

I would say that it sets the KFB to whatever value of the successfully obtained segment or 0 if no segment was obtained - you do not want a KFB with values relating to a previous call, would you?

Of course IMS keeps track of where it is - how is it possibly going to efficienlty keep track of where it is for the next call?

Elementary programming - and IMS is "just" another program or series of programs.

by **Ed Goodman** » Thu Aug 11, 2011 9:14 pm

The KFB can be changed for an unsuccessful call. You would see this if you are using more than one SSA.
For instance a root, first level child, and second level child. If the root and first level were found, but not the second level, the KFB would be set to the concatenated key of the root plus the first level segment. The length set to the length of those two keys.

by **Ed Goodman** » Thu Aug 11, 2011 9:18 pm

Since your GU has an EQ on a fully qualified key, IMS might, repeat MIGHT, be moving to the portion of the database where that key would show up according to the randomizer. I didn't see if you were using an HDAM or a HIDAM database.

Also, the KFB is NOT cleared before or after a call. That linkage area may still have old keys in it no matter what. That's why it's critical to check the KFB length.

by **dick scherrer** » Thu Aug 11, 2011 11:39 pm

Hello,

If you do not ensure you got "something" and try to use it anyway, beware of the very old caution from IBM.

The results may be unpredictable.

by **Mike1304** » Sat Aug 13, 2011 12:09 pm

I still believe, that my last posting shows the truth!

But I want to add, that if I test the same constallation with DFSDDLT0, i get a KFB after a not successfull GU!

by **Ed Goodman** » Sat Aug 13, 2011 9:36 pm

We believe you!

For the DFSDDLT0 KFB, what is the length? Can you post the results here?

by **dick scherrer** » Sun Aug 14, 2011 12:16 am

Hello,

i get a KFB after a not successfull GU!

Are you sure you "get one" or is this just something unpredictable or left over from the prior request. . . ?

by **Mike1304** » Mon Aug 29, 2011 5:59 pm

sorry for the late answer, but I didn't expect any new answers...
here the output from DFSDDLT0 PRINTDD

Select all

********************************* TOP OF DATA *************************************************************************************
************************************************************************************************************************

IMS 1210 TEST PROGRAM OUTPUT ** BEGIN TEST ****TIME= 14.21.19.26 DATE= 11.241

STATUS INPUT: S 1 1 1 1 1 1BVDU 00162003
0002 OF 0005 PCB SELECTED = BVDU DBD SELECTED = BVDU
........................................................................................................................
CALL=GU SEG=DUSGROOT FIELD=DUSCL OPER== VALUE=123456789012345678901234567890123456 00163003
7901234567890123)

NO SEGMENT RETURNED.

DBPCB LEV=00 SEG= RET CODE=GE KFDB LEN=0043 KEY FDB=(1234567890123456789012345678901234567901234)

*******************OSAM SUBPOOL STATISTICS******************************************************************************

B U F F E R H A N D L E R / O S A M S T A T I S T I C S STG CLS=00 FIXOPT=N/N POOLID:

IBM Mainframe Forum

key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU

Re: key feedback area after unsuccessful GU