[Info-vax] Raxco VMS Tuning Seminar Notes

[AEF]

On Nov 8, 3:06 pm, jls <notva... at yahoo.com> wrote:
> On 3 Nov 2010 16:33:45 -0600, koeh... at eisner.nospam.encompasserve.org
>
> (Bob Koehler) wrote:
> >In article <gr73d61ue65tf9uiv65fhdu6mg8ihqa... at 4ax.com>, jls <notva... at yahoo.com> writes:
>
> >> PFRATL does not trim idle processes.  It is a quantum-driven memory
> >> management technique and idle processes do not reach quantum end.
>
> >   So they eithre got trimmed before they went idle, or when my process
> >   watcher called $GETJPI, which queues an AST to them.
>
> If they get paged down to swpoutpgcnt, then when they become active
> again, they will have to page in each page separately, and if these
> are hard faults, it will be one I/O for each page.

I recall that the manuals say that swapping is a very expensive
operation I/O wise. Am I missing something here?

>
> And if you're using other mechanisms to reclaim memory, why use
> PFRATL?  The interaction between PFRATL and PFRATH almost insures that
> active processes will thrash pages in and out.  That does not help
> performance.

That's why there's

3.7 Active Memory Reclamation from Idle Processes

The memory management subsystem includes a policy that actively
reclaims memory from inactive processes when a deficit is first
detected but before the memory resource is depleted.

The active memory reclamation policy acts on two types of idle
processes:

    * Long-waiting processes
    * Periodically waking processes

[...]

AEF