DSXchange

When I open Operations Console the CPU Usage chart (lower left corner) always shows 100%... If I run topas it shows a very minimal usage, 1-35%... but not 100%.

I have installed a patch to allow me to change the Work Load Manager (WLM) parameters (I couldn't change anything before)... I did not try to use Ops Console before installing the WLM patch.

I engaged IBM Support on this issue two weeks ago.... and while they've asked for a lot of logs and such... no one has really had me do any troubleshooting. I looked at the Ops Console database in the ResourceUsage table... There are 15 columns in there named "CPU..." something... All of them are either 0.0 or NULL.

I have not discovered which process I have to stop to allow the data files in the ../DSODB/events directory to hang around... thinking I would look at the data in the files. Anyone have any ideas?

Does anyone have any other ideas on the cause of this issue? Or what I might do to have Ops Console correctly display the CPU%?

I appreciate your thoughts and any help you can give me.

Tony

For comparison, I'm running 11.3.1.0 with no patches, on AIX 7.1, and the CPU% bounces around and looks realistic. It may be tricky to match it up with topas due to timing--intervals and averages. In topas, add User%, Kern%, and Wait% together.

Perhaps the bug is on your AIX side and could use an AIX patch or new TL; check with an AIX Admin. I recall an in depth IBM developerWorks article about the Ops Console. You may want to search on that because I think it talked about which system calls it used.

On the Info Server side, you could try requesting that Support schedule a web session where they have the right developer involved to see it first hand.

Also ask Support if there are any other customers who have ever reported this issue. I don't know about you, but often times I get this feeling that Support is trained to make me feel like I'm the first and only one to report every issue!

Have you checked IBM Fix Central to see if there are any other patches available?

Eric, Thanks for your response.

I'm not sure how I'd get support to get the right guy on a web session. Support says they checked their problem system and said no one else has had this problem. I relayed my observations about the values in the database and complained that no one was working with me to troubleshoot the issue... and he said that since no one has had this problem all he could do was escalate this problem to Engineering, which he did right away. It seems to me that there are things we could be looking at to isolate this issue, but my support person doesn't seem willing or able to do any of this.

Good idea about looking in Developer Works. I'll do that.

It sounds like Support did the right thing. Perhaps the engineer you need is on vacation or is busy working issues affecting multiple customers.

Here is the article link. It tells exactly how the stats are calculated per operating system and includes some queries too.

Article: Interpret IBM InfoSphere DataStage and QualityStage Operations Console metrics

Wondering if you ever got a resolution to this one... customer on AIX is also seeing similar behavior.

Tony, did you ever get this fixed? I'm seeing 100% as well on my AIX setup.

Thanks in advance.