Subject: pmax memory requirements
To: None <port-pmax@NetBSD.ORG>
From: James MacKinnon <jmack@Phys.UAlberta.Ca>
List: port-pmax
Date: 04/29/1995 10:25:35
earlier last week, with a DS2100,

(machine with 8MB only will ultimately hang on first pageout to disk/swap,
 which prevented a kernel build).

After sticking another 16 MB of simms in the machine (24 total), it 
allowed me to build a kernel. The average memory usage during the
build hovers around 13-14 MB (as seen by vmstat), so even 16 MB is kind
of marginal.

This would seem to indicate that swapping to disk is non functional yet, 
even though this new kernel (which boots fine) defines:

config          netbsd root on rz0a swap on rz0b dumps on rz0b

SUP sources are from early April, and sysctl yields:

kern.ostype = NetBSD
kern.osrelease = 1.0A
kern.osrevision = 199306

It looks as if there is memory leakage too. After being up for about
24 hours, vmstat now shows depleted memory totals (I've lost about 10 MB):

dene# vmstat
 procs   memory     page                    disks         faults      cpu
 r b w   avm   fre  flt  re  pi  po  fr  sr r0 ?1 ?2 ?3   in   sy  cs us sy id
 0 0 0  5448  4704    8   3   0   0   0   0  1  0  0  0   69   71   6  5  4 92

at boot time this WAS approx:
 procs   memory     page                    disks         faults      cpu
 r b w   avm   fre  flt  re  pi  po  fr  sr r0 ?1 ?2 ?3   in   sy  cs us sy id
 1 0 0 11944 10456   86  24   2   0   0   0 11  0  0  0  555   74  37 49 18 33

vnode total has also grown from BOOT (and may be responsible for the 
memory depletion):

dene# pstat -T
 63/2588 files
   568 vnodes
0M/64M swap space

to(NOW):
dene# pstat -T
 63/2588 files
   2120 vnodes
0M/64M swap space


Does anyone know why page and swap is not cycling on the root's b partition
(now obvious from the 8MB behavior)?

pstat does say that it's there (even when I only had 8MB ram):
dene# pstat -s
Device      1K-blocks     Used    Avail Capacity  Type
/dev/rz0b       65536        4    65532     0%    Interleaved

I've been looking through the kern/ and bootcode sources, but nothing 
jumps out at me as being obviously inappropriate.

Cheers,
--
James S. MacKinnon           Office: P-139 Avahd-Bhatia Physics Lab
Department of Physics        Voice : (403) 492-8226
University of Alberta        email : Jim.MacKinnon@Phys.UAlberta.CA
Edmonton, Canada T6G 2N5
        WWW:   http://www.phys.ualberta.ca/~jmack/jmack.html