On Fri, 8 Nov 2019 at 12:59, van Sleeuwen, Berry <[email protected]>
wrote:

> Well, the affected LPAR is capped at 6% CPU. And as it turns out a SMAPI
> process is active at 5% for one or two minutes, and just in that minute the
> MONWRITE machine might have had just a fraction too little CPU time. Oddly
> enough we also have a reporting machine processing MONITOR records. Where
> MONWRITE simply writes selected records, the reporting machine has some
> serious plumbing but that machine had processed all records within time.
> (as for CPU usage, MONWRITE uses 0.07% CPU and MONREPT 0.20%, both have a
> share ABS 3%.)
>

If you haven't yet, get the number of logical CPUs down. A problem with
very "thin" CPUs is that z/VM is dispatching for a period in wall clock
time, but when you're dispatched only 6% of the time, it's likely that your
virtual machine gets empty time slices (the moment LPAR dispatches the
logical CPU, the time slice has ended already). When that happens a few
times, CP puts you in Q3 and gives you larger empty time slices even less
frequently. Your ABS 3% may not be worth a lot as that would be based on
when all logical CPUs get 100% each. But at least you have the monitor data
to figure it out.
I think z/OS recommends not to give a logical CPU less than 10% of a CPU
worth of weight; such rules of thumb should resonate with your capacity
planning friends as well.

Rob

Reply via email to