Re: Odd pool assignments

'Differing characteristics' is how I always looked at it. FOr example Interactive tasks typically sat Idle a fair bit. When the user got back to it, a big pause while it gets paged back in is bad. Batch jobs on the other hand had a tendency to use available memory and if in the same pool, elbow the interactive jobs out.

Today it's often SQL, loves to use what's available and that is logical. But whether it's Java, SQL, web servers, interactive jobs. There are clearly jobs that shouldn't be in the same pool. But it needs to be done by someone who understands with workloads and their characteristics.

- L

On 2/2/2021 11:21 AM, Joe Pluta wrote:

It's been a little bit since I messed with this, but if I recall one of the primary uses for separate pools was to separate Java VM load from "standard" IBM i OS load. The paging for the two approaches seemed to be incompatible, so separating them whenever possible seemed to be helpful.  Not sure that's a thing these days.  :)

On 2/2/2021 8:26 AM, Larry "DrFranken" Bolhuis wrote:

Well that's the flip. How many systems have I seen with 10 or more separate pools? Some private some shared often several that are not used at all! Yes there are valid reason for separate pools but these days not a much as in the past.

 - L

On 2/2/2021 7:41 AM, Rob Berendt wrote:

Good thought Larry.

Any chance they've went gonzo in breaking out stuff normally in *base into different subsystems and pools?

Rob Berendt