MVSFORUMS.com

satyenderd · Beginner Joined: 26 Aug 2005 Posts: 144 Topics: 73

Hi,
Let us explain the scenario:

Appln.team is running few batch jobs and the jobs are taking time around 7 hrs to complete. Previously, these jobs used to complete within 2 hrs.
We checked the MSTR and identified few tablespaces which were causing the contentions to other tables. In this paticular table it was accessing other partition tables throwing -911 errors.

I would like to know how to improve the performance of the jobs which runs very slowly. I have re-orged few tables which i found causing the contention problems. Later appln team compained the same problem.
Could any one please let me know the solution for this.

Thanks in advance.
_________________
Satya

warp5 · Posted: Tue Oct 30, 2007 7:23 am Post subject:

Are the programs using indexes for the tables or are they reading sequentially? Quite often performance problems in our house are caused by programs not using the indexes designed or they need a new index designed for the application.

jsharon1248 · Posted: Tue Oct 30, 2007 12:07 pm Post subject:

I assume that you ran RUNSTATS after the REORGs.

You're request is too vague. Nobody can do anything but guess based on what you've provided. The 'problem' could be growth of data. It could be contention for DB resources or a DB configuration change. It could be poor performing SQL. The DBA's need to monitor this process to isolate the problem. You need to tell us more about the problem. Then, we can recommend a course of action.

satyenderd · Beginner Joined: 26 Aug 2005 Posts: 144 Topics: 73

Thanks all for ur suggestions.
_________________
Satya

RMi · Posted: Wed Oct 31, 2007 7:32 am Post subject:

Satyender,

Since you said that the jobs used to take only 2 hours previously and now it takes 7 hours, I guess that there could be some new jobs introduced which are accessing the DB2 tables at the same time. It thats the case, you have to analyse them and reschedue the jobs appropriately.

Also the growth of data can be a reason for -911s. You can plan to archive the 'obsolete' data to keep the database as thinner as you can. This would help.

satyenderd · Beginner Joined: 26 Aug 2005 Posts: 144 Topics: 73

RMi

Thanks a lot for u r suggestion.
_________________
Satya

videlord · Beginner Joined: 09 Dec 2004 Posts: 147 Topics: 19

for performace problem, I think you'd better get DB2PM Accounting report for the plan (batch job)
checking MSTR log can't help

and you should check what has been changed (2 hours -> 7 hours)
Hardware? OS or software version? system parameter? application program? table structure? data in the table?