View previous topic :: View next topic
|
Author |
Message |
saikiran A
New User
Joined: 23 Aug 2012 Posts: 2 Location: India
|
|
|
|
Hi,
We recently converted a DLI batch program to BMP. The job started abending with U0775. After initial research, we found that PIMAX was being exceeded. We started off with a checkpoint once every 2000 records and finally settled at 2. But, the U0775 continued.
Later, by looking at DFSTAT, we realized that the child to parent ration of the IMS database (which is the driver for pgm) is around 1500. All these child segments were being enqueued due to a PROCOPT of A. This huge enqueue of childs + root data is probably crossing the PIMAX.
IMS Locking is Program Isolation. Please suggest if there is a way to avoid the huge/enqueue while still having PROCOPT=A. PROCOPT=A was needed since DLETs are required on the database. |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
ENQ is there to protect Your data...
speak to Your support to raise the limits of the ENQ buffer pool and friends |
|
Back to top |
|
|
saikiran A
New User
Joined: 23 Aug 2012 Posts: 2 Location: India
|
|
|
|
Thanks Enrico,
DBA are already in the process of increasing PIMAX.
I am just curious if PROCOPT or something else can be adjusted to achieve the DELETEs without ENQ of the whole IMS record. something similar to ENQ of root alone.
One other detail is that the job can have exclusive access of the database. No other job or online user is expected to access the database while the job runs. |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
Quote: |
No other job or online user is expected to access the database while the job runs. |
that' the usual horse manure ( that You were told )
be sure that sooner or later the same person will tell that You did not understand the requirement to it' s full extent |
|
Back to top |
|
|
|