[Info-vax] vms base priority watch

[Paul Sture]

In article <bJImDVCcxL7D at eisner.encompasserve.org>,
 koehler at eisner.nospam.encompasserve.org (Bob Koehler) wrote:

> In article <00AB2409.145ED576 at SendSpamHere.ORG>,   VAXman-  @SendSpamHere.ORG 
> writes:
> > 
> > I still do not see how report generation is all that CPU intensive that an
> > increase in a user's base priority would aid its completion time.
> 
>    It could be a bit CPU intensive, and if the system is loaded 100% all
>    day as the OP said, it could be loaded in the wee hours of the morning
>    enough to make CPU access a real issue.

An analysis of the business requirements seems appropriate here.  If for 
example it's the same reports every day which have to have their 
priority boosted, so that for example they are available by 9 am, they 
could (perhaps) be started with a higher priority in the first place.

If OTOH the priority boosts are triggered by customers calling up, then 
this isn't appropriate.

>    Ever since VMS 4.0, multiple I/O requests to the same device are 
>    prioritized by the base priority of the requesting process, so if 
>    there's a lot of contention for the disk holding the data or the 
>    report, that could impact the completion time.  I recall processing 
>    that was affected that way, the affected users reported it as a bug, 
>    and thier boss intercepted the bug report because he liked it that 
>    way (it was he who got them the lower base priority in the first place).
> 
>    He really did have higher priority stuff going on, and had thought
>    about suspending thier processes from time to time, but we never
>    got that far.

Indeed I recall having to do some "juggling" to get work through in 
pre-V4.0 days.

-- 
Paul Sture