We used a different approach to this. Our CPU is a quad. We do not run
VSE's turbo dispatcher in multi-engine mode as little of our workload is
able to benefit. This limits each VSE guest to dispatch on a single CP and
minimizes one VSE's ability to dominate the entire box. Note that each
guest's workload "fits" MIP-wise onto a single CP, the guests do not sit at
100%.
We have multiple production guests and balance the work across them as
evenly as possible.
We also use prty share within the balanced partitions, with the lowest value
being 100 and the highest being 3000 or more. Weird? Probably. But quite
effective in our environment with our workload.
Mike Horlick wrote:
I'm curious how you are able to run in one VSE machine with CICS/TS
running with a relatively heavy batch job and not have it affect other
VSEs if pushed to 100% CPU utilization?
Thanks,
Mike
