gsg wrote:
We have a job that has a posting step that runs 6-7 minutes last year. We noticed that the job more than doubled starting with the first processing day of 2008. The application owners swear that no changes were made on there side and the systems guys swear that no changes were made on their side neither. The only think that I noticed by looking at the jobs sysout is that the SERVice Units went from 5335K to 13549K. Any ideas.
Run time is NOT a good indicator of performance issue, unless absolutely NOTHING in the WHOLE environment has changed. By nothing I mean NOTHING.
The same exact jobs are running at the same exact time this job is running. This job is processing the same exact data. You have the same number of online users doing the same exact work.
Say this job was taking 6-7 minutes "last year". Now on Jan. 1 of this year a new application was put into production and is consuming more resources (CPU, MEMORY, I/O) so that your whole system is busier than last year. This job could be running longer (walk clock time) because the system is busier and this job has to wait more often and longer to get the resources it needs.
IIRC service units can and do change from run to run. The biggest reason is the number of records processed by the job.
I agree with Arhtur, look at the link edit dates. The hard part about this, if your shop is like ours, we have a few common modules that are dynamically linked so the "main" program may not have changed, but one of the common modules did and so you have to check all of the modules that are dynamically linked also.
---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
