Can anyone please let me know if we can use COMMIT calls in IMS/DB2 programs.
I have an IMS/DB2 program that does not have any check point re-start logic. It reads an input file, checks few data in IMS database and updates the DB2 table. Wanted to introduce a COMMIT call after each 10,000 records updated in the table. Will this work out?
Or do I need to go for a checkpoint re-start logic itself?
Joined: 23 Nov 2006 Posts: 19270 Location: Inside the Matrix
Adding COMMITs implies that checkpoint/restart logic is also needed. . .
Suggest you re-examine the desire to add checkpoints - many systems rerun from the beginning of the job as this is often cleaner/faster than trying to restart "in the middle" (unless the restart operation is already in place and works smoothly).
The faster cpus and dasd make a full rerun a better choice than restart/recovery for most jobs.
When used, restart/recovery design should not be done at the job level, but rather at the system level so everything is consistent.