Unicenter 3.0 NSM Workload Management Option 1

[rstinnett]

Howdy -

I've been given the daunting task of integrating Unicenter 3.0 NSM Workload Management into our organization; we also purchased OpenVMS Console Manager, JobVision and OpenVMS Workload Manager.

My background with Unicenter has all been mainframe (CA-7, CA-11, etc.) but I'm the only guy who knows what the heck these products do, so I've been assigned to head up this project.

Many questions abound; and hopefully some of you can fill in some of the missing pieces of my puzzle.

First of all; What is the difference between Unicenter 3.0 and Unicenter 3.0 Workload Management Option? Their support pages only talk about the "full blown" version and NOWHERE do they give the slightest clue as to what Workload Manager has/doesn't have compared to the full product.

Second, when scheduling jobs, does Unicenter have the ability to go "Oh, I see your CPU is currently at 90%, so I am not going to throttle your box until it does". Or, furthermore, can we setup jobs that say "Only kickoff this job when the CPU is below 75%".

So far, I am NOT impressed with the documentation that Unicenter has. The manual is so vague, it doesn't touch on anything specific.

We have a consultant scheduled to come in here for two weeks in December and get the ball rolling (that is if CA can find one for us!). However, I don't believe in waiting until the last minute to figure some of this stuff out and I certainly don't believe in relying on a consultant for all of our knowledge.

 

[JonnoGT]

CA have packaged everything differently with Unicenter 3.0
So when you buy any option you get the "common services" with it. This is the equivialent of the old framework - world view and event. So when they talk about Unicenter 3 on its own I think they mean the common services. Workload Management option will include all of this.

I have been using TNG 2.1 and 2.4 workload for years now and I do not think there is much different in version 3.0. There is no direct way of starting a job off based on CPU usage, unless you were to fire off jobs from event management based on agent messages. This is pretty easy to do, but from a batch design point of view, I wouldn't recommend it.

You do have the ability to give jobs shared resources, so you can ensure that no more than X amount run on a box at any one time.
And triggers - so that a job will not start until a file has arrived etc.
And of course standard job dependancies (dont run jobB until jobA has finished etc).

You'll find most of the useful documentation in the Unicenter Implementation CD (available from their web site).
Or do a find on CAISCHD in the reference guide. This will tell you all of the environment variables / settings you can change for workload.
Other than that, there isn't much else. Frustrating isn't it?