Page 1 of 1
Performance with 7.1r2
Posted: Fri Jan 19, 2007 2:50 am
by thurmy34
Hi
Since we migrate to the 7.1r2 from the 7.1r1 we have some issues.
We have job that run forerever.
We have job that last 2:44 minutes r1 and now last 47:28 minutes.
Did someone encounter the sames troubles ?
Thanks for Helping
Posted: Fri Jan 19, 2007 4:06 am
by ArndW
No, I haven't seen anything like that. What does your job do? Did you change anything else at the same time (OS or patch level, DB version, etc.)?
Posted: Fri Jan 19, 2007 4:48 am
by thurmy34
Hi,
My job split rows and use a link collector before writing in a DB2 database.
I'll investigate if others changed where made with this upgrade.
Posted: Fri Jan 19, 2007 4:56 am
by ArndW
If you write to a sequential file instead of DB2 does it speed up greatly? (it should, if not you do have some issue). Do other jobs that write to DB2 also show a slowdown?
Posted: Fri Jan 19, 2007 5:26 am
by ray.wurlod
Do you really need a Link Collector? In theory your job should go faster with parallel independent streams of inserts.
Posted: Mon Jan 22, 2007 3:43 am
by thurmy34
Hi,
I'm sorry for the delay.
I changed my job by deleting the link collector and using a pivot stage.
The job runs with the time it had in the previous version.
I'm just worried about the behavior of Datastage. A minor change shoudn't have a huge impact.
I have to change a lot of jobs , because they hang in r2 plus the solution that the support gave me works in dev but not in production.
Thanks for helping.
Posted: Mon Jan 22, 2007 4:16 am
by ray.wurlod
Any version number containing "r" is a "rapid repair fix" and has not been through full Quality Assurance testing. You should contact your support provider for a version that has been through full QA testing.
Posted: Tue Jan 23, 2007 8:16 am
by shrek7577
Hi,
I've the same problem toda...
viewtopic.php?t=106909
"job running but no log, no PID, too long, CPU Pb ?"
Big coincidence !
Posted: Tue Jan 23, 2007 8:56 am
by ArndW
shrek7577 - it doesn't look like quite the same problem to me, this is a n extreme slowdown type of issue while what you described was quite the opposite, where the job doesn't terminate correctly.