WML impact on performance

A forum for discussing DataStage<sup>®</sup> basics. If you're not sure where your question goes, start here.

Moderators: chulett, rschirm, roy

Post Reply
thurmy34
Premium Member
Premium Member
Posts: 198
Joined: Fri Mar 31, 2006 8:27 am
Location: Paris

WML impact on performance

Post by thurmy34 »

Hi
I would like to konw if it's possible to see that a job has been queued before its start ?
We have loss of performance with the version 11.5 (from the 8.7) and i want to mesure the wml impact.
Thank you.
Hope This Helps
Regards
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

You can see it in real time from Director as the jobs are run. The status will say Queued instead of Running.

You can also see it in near real time from the Operations Console. Go to the Workload Management (WLM) tab. In the top box under System Policies you will see some boxes light up bright red if the settings currently trigger jobs to queue.

in the lower section on the same tab, select the Queued Jobs tab. Set a useful refresh interval. There you can watch to see which specific jobs are waiting in queue.

If you want to look at the history of queued jobs or how much time they spend waiting in the queue to start, presumably to protect your server from overload conditions, then I am not sure if that is possible. If not, just watch the console during active times.

If you want to see for sure if WLM is overall helping or harming then you could disable it for a day or a week and measure the impact that way. I would just caution that if you disable it, to be careful and monitor closely that you don't allow the jobs to use up all the memory including all the virtual memory. There is a separate topic here somewhere that tells how to disable WLM.

You could also effectively nearly disable WLM by changing all the system policy settings.
Choose a job you love, and you will never have to work a day in your life. - Confucius
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

WLM is not necessarily the culprit; it has been designed to be as lightweight as possible.

However, its threshold settings are, shall we say, conservative? It won't allow jobs to start if the idle CPU is under 20% or the available memory is under 20%. If you tune these "closer to the bone" then fewer jobs will queue, and are likely to spend less time queued. On the other hand, the risk of overloading your server, which you possibly were doing in version 8.5, is commensurately increased.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Post Reply