MVSFORUMS.com

[arshadh]

Dear All,

In our database we have segmented tablespaces (i.e The table spaces are partitioned based on the Product type the partition can hold..)

Now the Problem is when we run a maintenance job for two different products (which means data is accessed from different partitions ) one of the job fails due to contention problem

[videlord]

1. partitioned table space is not a segmented tablespace
2. REASON 00C9008E is timeout, not dead lock
3. TYPE 210 is resource partition
The batch job PRDNCDLY and PRDNJDLY must require lock on the same partition PRD0001D.PRD0010S.00000001

check the program and table parition difinition

[arshadh]

Thanks for ur comments..

Can you explain in detail the things I need to done..?I really can't understand your third and fourth point...

Our table space is divided based on the products. Both PRDNCDLY and PRDNJDLY are two differnet products which for sure wont fight for a same page in the TS...

Pls throw some light on this...

[videlord]

based on the message, the 2 jobs access the same partition.
The lock level is partition, not page.

Please give detail information:
1, the partition index DDL (index key range)
2. The SQL in the 2 jobs
3. The input data to the job (if the SQL execution depend on the input)

[jsharon1248]

Resource Type 210 is an index space. You're most likely updating columns that are defined to indexes that are not the partitioning index. What is your commit frequency? You might be able to resolve this by issuing commits more frequently to release the index locks.

I seriously doubt that the SORT would in any way cause the DB2 timeout, unless you're running the SORT within a unit of work which would increase the duration of the unit of work. But that's really not the SORT causing the problem, it's a design issue.

[videlord]

[jsharon1248]

Dyslexai strikes again....

[arshadh]

Thanks for your Suggestions....