At Wed, 16 Nov 2011 18:13:34 +0000 (UTC), Eduardo Horvath
<eeh%NetBSD.org@localhost> wrote:
Subject: Re: non-automated test failure report! :)
>
> Frankly I disagree. If you want a robust, stable production system you
> want to not overcommit memory. You only want to turn the checks off if
> you don't care about the stability of the machine. If you want to play on
> the wild side then you might want to enable overcommit by a certain
> percentage. But now you don't have that option. The machine runs fine
> until it falls over.
Indeed -- I'd love to have a switch to turn off over-commit entirely,
and one that permitted a slight amount under certain circumstances would
also be very nice to have.
Combined with this the ability to dynamically create additional swap
files in a filesystem would also be nice -- then one could also define a
policy that says the VM is allowed to allocate some amount and/or ratio
of space on a given (set of) filesystem(s).
I personally would also like to have some sort of mechanism to
automatically deal with wildly growing page-out rates. If some process
or set of processes are growing at a rate which is about to drive the
system into thrashing with paging activity then it would be nice to be
able to define a policy about what to do. For example if it's a certain
class of process, or class of user, etc., then maybe it's killed,
warned, or maybe it's just stopped, etc.
--
Greg A. Woods
Planix, Inc.
<woods%planix.com@localhost> +1 250 762-7675 http://www.planix.com/
Attachment:
pgpdRciDCgOBG.pgp
Description: PGP signature