|
View previous topic :: View next topic
|
| Author |
Message |
sherlinpraveenkumar
New User
Joined: 14 Jul 2010
Posts: 6
Location: chennai
|
|
|
|
A build alternate index step fails while executing a JCL due to the following error message and throws RC=04.
Build alternate index steps fails due to QI309 INSUFFICIENT "AIX" RECORD LENGTH - FOLLOWING BASE KEY(S)/RBA(S) DROPPED
Note: This job is daily job and it does not fail regularly.
Follwoing is the DEFINITION of the AIX:
DEFINE AIX (NAME(TTT.XXX.ZZZ.AIXINV)
RELATE(TTT.XXX.ZZZ)
VOLUMES(PRD136)
KEYS(25 55)
RECORDSIZE(900 670000)
CYLINDERS(10 1)
NONUNIQUEKEY
SPEED IMBED)
DATA
(NAME(TTT.XXX.ZZZ.AIXINV.DATA)
FREESPACE(20 20)
CISZ(4096))
INDEX
(NAME(TTT.XXX.ZZZ.AIXINV.INDEX))
DEFINE PATH (NAME(TTT.XXX.ZZZ.PATHINV)
PATHENTRY(TTT.XXX.ZZZ.AIXINV))
I hve gone thru the reference link, but still would like some assistance in understanding and fixing it. Can someone help me? |
|
| Back to top |
|
 |
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
| You have duplicate alternate index keys. Each existing key is a record with the primary key of each record which matches that key. You now, today, have more duplicates for a particular value (at least one) than previously, and that happens to make your alternate index record too small. Make it bigger. If it is already maximum, then you have thinking to do which should have been done at the design stage. How did this get through testing? |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
| Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009
Posts: 580
Location: London
|
|
|
|
It would help if you showed the BASE CLUSTER define statement.
The QI309 message is also unfamiliar, are you using a 3rd party product of basic IDCAMS to do the define?
Not sure how you can have RECORDSIZE(900 670000), i.e. a 670k recordsize!? I think the record size is limited to 32760 bytes. Also the CISIZE is only 4096 bytes and you haven't specified SPANNED so even a 32k record size wouldn't fit.
Suggest you read the IBM manuals 'VSAM Demystified' and 'Using Data Sets' urgently. |
|
| Back to top |
|
|
|
|
