Subject: Re: Bad response...
To: Julio M. Merino Vidal <jmmv@menta.net>
From: The Grey Wolf <greywolf@starwolf.com>
List: current-users
Date: 08/31/2004 08:23:10
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thus spake Julio M. Merino Vidal ("JMMV> ") sometime Today...

[I still keep crossing (JMMV, "Joor Mileage May Vary").  Sorry, Julio!]

JMMV> I've also seen this, and IIRC, it started this January more or less.
JMMV> For example, my server has 80 mb of ram, and it is quite loaded.
JMMV> It performed very well with NetBSD 1.6, and it hardly ever touched
JMMV> the swap.  However, just after updating it to NetBSD 2.0_BETA (when
JMMV> the branch was first cut), running the same services and the same
JMMV> versions of all packages, it started swapping.  ATM, I see it has
JMMV> used 10 mb of swap (not so bad, and I can accept it because of its
JMMV> load).
JMMV>
JMMV> However, my workstation exposes the problem more visibly.  It's an
JMMV> Athlon XP 2600+ with 512 mb of ram, running a very recent -current.
JMMV> Building a release and using GNOME or KDE in the meantime makes it
JMMV> swap a lot (specially if you go away and leave the machine unused
JMMV> for several minutes), up to 120 mb or so (it was worse at the
JMMV> beginning of the year).

Let me toss in a bit of a wrench.  Or not.  This is simply a data point
to be included and analysed.

Hardware:
Athlon XP 1900+, 512MB core, DMA 4/5 disks, most all else irrelevant.

Kernel:
NetBSD rivendell.starwolf.com 2.0G NetBSD 2.0G (RIVENDELL) #0: \
	Sat Aug 14 12:19:11 PDT 2004 \
	greywolf@rivendell.starwolf.com:\
	/.a/src/sys/arch/i386/compile/RIVENDELL i386

Userland is from 16 july 2004.

$ sysctl vm | egrep -e 'min|max'
vm.anonmin = 10
vm.execmin = 5
vm.filemin = 10
vm.maxslp = 20
vm.anonmax = 80
vm.execmax = 30
vm.filemax = 50
$

No problems here compiling the world and listening to music.  Runs VERY
smoothly.

I've only had the occasional reboot while doing stupid things like
running backups and building things on the same filesystem :), and once
the audio driver for my s3 didn't attach properly so that every time
I started xmms, the kernel panicked immediately.  A rebuild seemed
to (mysteeeeeriously) fix that problem.

I'm about to upgrade to another -current release; if anything goes wrong
or different, I'll post here.

About the only thing I notice is that a build-world now takes 1:10:00,
roughly, where it used to take 0:53:00 about a year ago, and where I
was able to build the world in 1:05:00 on a P3/500 in 1999...

				--*greywolf;
- --
[ ] PEACE    [ ] FREEDOM	Pick one.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFBNJfiDOGizqAnVRERAsdLAJ96AZ1GaooRng4M6EB823NTXHPVWACgk+qM
JNjeOM1S2kqgdY4Det124pI=
=IE87
-----END PGP SIGNATURE-----