Date: Mon, 18 Sep 2000 00:01:32 +1100 (EST) From: =?iso-8859-1?q?Paul=20Jansen?= <vlaero@yahoo.com.au> To: hackers@freebsd.org, questions@freebsd.org Subject: diskless - /sbin/init won't start over NFS mounted root Message-ID: <20000917130132.18895.qmail@web5103.mail.yahoo.com>
next in thread | raw e-mail | index | archive | help
I'm having troubles getting a diskless system to boot
using a FreeBSD 4.1R server and a custom compiled
FreeBSD 4.1R kernel. I've made sure the following
options were defined in my custom kernel:
options NFS_ROOT #NFS usable as root
# Kernel BOOTP support
options BOOTP # Use BOOTP to obtain
IP address/hostname
options BOOTP_NFSROOT # NFS mount root
filesystem using BOOTP info
options BOOTP_NFSV3 # Use NFS v3 to NFS
mount root
options BOOTP_COMPAT # Workaround for
broken bootp daemons.
I'm using the ISC v2 DHCP server. Here's the
/usr/local/etc/dhcpd.conf file:
# dhcpd.conf
# option definitions common to all supported
networks...
option domain-name "jansen.org";
option domain-name-servers 202.167.41.130;
option subnet-mask 255.255.255.0;
default-lease-time 600;
max-lease-time 7200;
subnet 192.168.0.0 netmask 255.255.255.0 {
range dynamic-bootp 192.168.0.10 192.168.0.20;
option routers 192.168.0.8;
option broadcast-address 192.168.0.255;
filename "/tftpboot/kernel.XWORKSTATION";
option root-path
"192.168.0.1:/usr/diskless/rootfs/xworkstation";
option option-128 "192.168.0.1:/usr/diskless/swap";
option option-132 00:00:08:00;
option option-176 5;
option option-160 "timeout=8:default=193:";
option option-192 "IJ2:::/tftpboot/kernel.ij2:";
option option-193
"XWORKSTATION:::/tftpboot/kernel.XWORKSTATION:";
}
The boot process get all the way down to mounting root
and swap over NFS. The line:
option option-132 00:00:08:00;
in the dhcpd.conf file above allows for verbose
information to be displayed to the console when
booting. The last few lines displayed on the console
boot screen are:
Mounting root from NFS:
NFS ROOT:192.168.0.1:/usr/diskless/rootfs/xworkstation
NFS SWAP:192.168.0.1:/usr/diskless/swap
start_init: trying /sbin/init
So the system has a problem running /sbin/init over
the nfs mounted root. I've had a look at what's
happening on the wire using ethereal. This is what it
turns up:
first of all the diskless system sends an NFS request
to the server of type 'LOOKUP CALL'. Ethereal> says
that proceure LOOKUP is happening in this packet.
The next packet goes from the server to the> diskless
machine. The procedure is again LOOKUP but this time
the packet also has an status entry for NFS. The
status says:
ERR_NOENT (2)
My /etc/exports file has the following line in it:
/usr -alldirs -maproot=0:0edessa pico host10 host11
host12 host13
I've checked that the filesystem referenced above is
being successfully exported using NFS by mounting it
on another machine on the network. I can go into the
/sbin directory and execute init. Of course the other
machine complains because it already has init running.
This just goes to show that the exports directory is
mountable and the files are executable.
Does anyone have any ideas as to what is wrong here.
I am stumped and am quite keen to get diskless
operation happening soon.
Thanks in advance,
Paul
_____________________________________________________________________________
http://geocities.yahoo.com.au - Yahoo! Australia & NZ GeoCities
- Build your own Web Site - for free!
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000917130132.18895.qmail>