View previous topic :: View next topic
|
Author |
Message |
supreethi.srid
New User
Joined: 15 Oct 2009 Posts: 5
|
|
|
|
Hi,
I am stuck with an abend - U0828 , am here to find a way to get rid of it.
I am trying to run an IMS DLI program that tries to insert / delete / replace segments based on values and validations from the input file. I am using two database clusters - one DB and one its index.
My program is able to process a few records without issues, but few records give me a status code of 'NI' while ISRT call.
I tried to bypass these records by evaluating PSBmask of ISRT call-status(in case of 'NI' status), but still error persists.
Quickref :
"Index Maintenance attempted to insert a new index entry, with index
specified as unique, and encountered a duplicate index entry.
If duplicate secondary index entries occur, the index should be specified
non-unique, and an overflow entry-sequence (EDS) data set should be
provided. If running batch, backout should be executed to resynchronize
data base indexes. "
I dont want to change the index segment as non-unique in the DBD.
Can someone let me know a solution to solve this abend?
Appreciate early response! |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
Quote: |
I dont want to change the index segment as non-unique in the DBD.
Can someone let me know a solution to solve this abend?
Appreciate early response! |
I wonder how You were able to get the NI status code
usually IIRC the abend is scheduled before letting You process the status code
options :
1) the one You do not want to implement
2a) do not insert anything that would cause the issue described ( before the damage )
2b) ugly commit after very succesfull operation
rollback and log the invalid update attempt ( after the damage )
anyway ) review the applications design/requirements/constraints
do not complain about the reply, Your issue was clearly described by the status code itself
nothing reasonable can be done about data/operations that violate the data base specifications/implementation |
|
Back to top |
|
|
supreethi.srid
New User
Joined: 15 Oct 2009 Posts: 5
|
|
|
|
Quote: |
usually IIRC the abend is scheduled before letting You process the status code |
Ha! Good Grief, I happened to suspect it !..
2a) seems to be good . Will try that ! Lets c.
Thank You Enrico ! |
|
Back to top |
|
|
supreethi.srid
New User
Joined: 15 Oct 2009 Posts: 5
|
|
|
|
Finally it happened to be an issue with my test database - that had 'pointer errors' as reported by DBAs...The DB was corrupted even before i ran my test ! Got the prod image copy extracts and checked - It ran fine !!!! |
|
Back to top |
|
|
|