If you get this kind of problem with MkLinux 2.1u5, and you have
a lot of RAM, read on:
*** SERVER IS KERNEL-LOADED ***
*** USING ACTIVATIONS ***
Mach 3.0 VERSION(PMK1.1): gilbert <osfmk>; Tue Oct 21 13:50:08 PDT 1997;
<lev 0> free_area_pager_init: vm_map err=0x3=3 "os/kern) no space
No debugger configured - dumping debug information
version string : Mach 3.0 VERSION(PMK1.1): gilbert <osfmk>; Tue Oct 21
13:50:08 PDT 1997; mach_kernel/PRODUCTION (venus)
bcktrace: 0021bf58 00230468 00214510 0020c0a0 002612af 00000000 00f27d00
100f4ea8 100cec44 100ceba8 100c99a8 10017ac4 100c335c 100008d0 100cfc6c
Please report the above messages to
firstname.lastname@example.org and reboot your machine.
It is important to set a flag in the lilo.conf file (accessible
through the MkLinux control panel) which prevents the kernel from
trying to allocate too much memory. Look at the README file
from the update directory, it tells you what to do if you
have a lot of memory.
(I think the above refers to the -m flag, see below)
> - A second segment register was added to the collocated server space. This
> means that the Linux server can map up to 512M when collocated.
> NOTE: for systems with more than 160M of RAM, you should change a line
> in your /mach_servers/bootstrap.conf file so that the Linux server
> doesn't grow bigger than it thinks it can handle. The line you will
> want to have is:
> -k -S 524288000 startup /dev/boot_device/mach_servers/vmlinux
> Make this change _before_ you remove the "-m" option from your lilo.conf,
> if you had used -m in the past.
FWIW, -k -S 524288000 startup /dev/boot_device/mach_servers/vmlinux
let me use u5 on my 7300/180 with 128M RAM. email@example.com