Installation problems with NFS mount on sun Fire X2100 with suspicious kernel behavior

Lukáš Oliva olivalukas at gmail.com
Thu Nov 19 20:20:26 CET 2009


  Hello all,
I used fai to install approximately 20 servers from Sun, but I am
completely unable to use it on Sun Fire X2100 machines (verified on
three of them). I use Debian Lenny on server (version 3.2.17~lenny1)
and I want it to be installed on all the machines. My problem is even
stranger, because fai seems to behave differently depending on kernel
I give the tftp to transfer.

The TFTP part works okay and transfers vmlinuz and initrd.img files,
then systems boot with no apparent error. Nevertheless when it comes
to the point where it should ask for mounting NFS, it behaves strange.

1) 2.6.26 (default lenny kernel image) will stop here and seems to do
nothing, it just waits. I can reboot it, even after hours, but it does
not ask for kernel not depending on the ip= parameter in approptiate
config.
2) 2.6.30 (kernel image from backports) kernel image ends with Kernel Panic.

I suspected ramdisk from causing problems, so I tried to compile
everything necessary directly into kernel (all nfs and nfsroot related
options, network card drivers and firmware) which resulted in ...

3) my own compiled 2.6.31 which is able to
  - waits quite long (longer than other machines) - during "sending
DHCP and RARP requests" phase
  - for some reason that I do not understand, it asks for IP address
in network 200.0.0.0, but gets NAK
  - take the correct IP address
  - ask for mount for nfsroot (authenticated mount request in
/var/log/daemon.log) which none of previous was able to do.

Unfortunately it does not continue with mounting the config directory,
but hangs on this message:

IP-config: Complete:
device=<ok>, addr=<ok>, mask=<ok>, gw=<ok>,
host=<ok>,nis-domain=(none), bootserver=<IP ok>, rootserver=<IP ok>,
rootpath=

I have no idea what wrong is happening here. Could anyone from
comunity help, please?

Here is what I consider to be relevant part of /var/log/daemon.log and
dhcp from syslog

Nov 19 19:57:28 x2100-01 dhcpd: DHCPDISCOVER from 00:1b:24:f0:2f:5c via eth2
Nov 19 19:57:28 x2100-01 dhcpd: DHCPOFFER on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 19:57:36 x2100-01 dhcpd: Dynamic and static leases present for
98.101.1.24.
Nov 19 19:57:36 x2100-01 dhcpd: Remove host declaration x2100-04-se or
remove 98.101.1.24
Nov 19 19:57:36 x2100-01 dhcpd: from the dynamic address pool for 98/8
Nov 19 19:57:36 x2100-01 dhcpd: DHCPREQUEST for 98.101.1.24
(98.101.1.12) from 00:1b:24:f0:2f:5c via eth2
Nov 19 19:57:36 x2100-01 dhcpd: DHCPACK on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 19:57:36 x2100-01 in.tftpd[27163]: RRQ from 98.101.1.24
filename pxelinux.0
Nov 19 19:57:36 x2100-01 in.tftpd[27163]: tftp: client does not accept options
Nov 19 19:57:36 x2100-01 in.tftpd[27164]: RRQ from 98.101.1.24
filename pxelinux.0
Nov 19 19:57:36 x2100-01 in.tftpd[27165]: RRQ from 98.101.1.24
filename pxelinux.cfg/3e253d95-f979-0010-afbb-001b24f032e9
Nov 19 19:57:36 x2100-01 in.tftpd[27165]: sending NAK (1, File not
found) to 98.101.1.24
Nov 19 19:57:36 x2100-01 in.tftpd[27166]: RRQ from 98.101.1.24
filename pxelinux.cfg/01-00-1b-24-f0-2f-5c
Nov 19 19:57:36 x2100-01 in.tftpd[27166]: sending NAK (1, File not
found) to 98.101.1.24
Nov 19 19:57:36 x2100-01 in.tftpd[27167]: RRQ from 98.101.1.24
filename pxelinux.cfg/62650118
Nov 19 19:57:36 x2100-01 in.tftpd[27168]: RRQ from 98.101.1.24
filename vmlinuz-2.6.31local-squashfs
Nov 19 19:57:37 x2100-01 in.tftpd[27169]: RRQ from 98.101.1.24
filename initrd.img-2.6.31local
Nov 19 19:59:08 x2100-01 dhcpd: DHCPREQUEST for 200.0.0.0 (62.0.0.0)
from 00:1b:24:f0:2f:5c via eth2: wrong network.
Nov 19 19:59:08 x2100-01 dhcpd: DHCPNAK on 200.0.0.0 to
00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:19 x2100-01 dhcpd: DHCPDISCOVER from 00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:19 x2100-01 dhcpd: DHCPOFFER on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:19 x2100-01 dhcpd: Dynamic and static leases present for
98.101.1.24.
Nov 19 20:00:19 x2100-01 dhcpd: Remove host declaration x2100-04-se or
remove 98.101.1.24
Nov 19 20:00:19 x2100-01 dhcpd: from the dynamic address pool for 98/8
Nov 19 20:00:19 x2100-01 dhcpd: DHCPREQUEST for 98.101.1.24
(98.101.1.12) from 00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:19 x2100-01 dhcpd: DHCPACK on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:20 x2100-01 dhcpd: DHCPDISCOVER from 00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:20 x2100-01 dhcpd: DHCPOFFER on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:20 x2100-01 dhcpd: Dynamic and static leases present for
98.101.1.24.
Nov 19 20:00:20 x2100-01 dhcpd: Remove host declaration x2100-04-se or
remove 98.101.1.24
Nov 19 20:00:20 x2100-01 dhcpd: from the dynamic address pool for 98/8
Nov 19 20:00:20 x2100-01 dhcpd: DHCPREQUEST for 98.101.1.24
(98.101.1.12) from 00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:20 x2100-01 dhcpd: DHCPACK on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:21 x2100-01 dhcpd: Dynamic and static leases present for
98.101.1.24.
Nov 19 20:00:21 x2100-01 dhcpd: Remove host declaration x2100-04-se or
remove 98.101.1.24
Nov 19 20:00:21 x2100-01 dhcpd: from the dynamic address pool for 98/8
Nov 19 20:00:21 x2100-01 dhcpd: DHCPREQUEST for 98.101.1.24
(98.101.1.12) from 00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:21 x2100-01 dhcpd: DHCPACK on 98.101.1.24 to
00:1b:24:f0:2f:5c via eth2
Nov 19 20:00:21 x2100-01 mountd[14691]: authenticated mount request
from 98.101.1.24:987 for /srv/fai/nfsroot (/srv/fai/nfsroot)

Here are relevant parts of /etc/dhcp3/dhcpd.conf:

authoritative;

allow bootp;
allow booting;

allow unknown-clients;
option dhcp-max-message-size 2048;
use-host-decl-names on;

log-facility syslog;

subnet 98.0.0.0 netmask 255.0.0.0 {
   range 98.101.1.1 98.101.1.200;
   option routers 98.101.1.12;
   option domain-name "mydomain";
   option domain-name-servers 98.101.1.12;
   option time-servers 98.101.1.12;
   option ntp-servers 98.101.1.12;
   server-name faiserver;
   next-server faiserver;
   filename "pxelinux.0";
}

..
# For each host I have ...
host x2100-04-se {hardware ethernet 00:1b:24:f0:2f:5c;fixed-address
x2100-04-se;}


Here is pxelinux configuration file:
default kernel-2.6.31local-reconfigured-squash

label fai-generated
kernel vmlinuz-2.6.26-2-amd64
append initrd=initrd.img-2.6.26-2-amd64 ip=dhcp  root=/dev/nfs
nfsroot=/srv/fai/nfsroot boot=live  FAI_FLAGS=verbose,sshd,reboot
FAI_ACTION=install console=tty1 console=ttyS1

label kernel-2.6.31local-reconfigured-squash
kernel vmlinuz-2.6.31local-squashfs
## ip=<client-IP-number>::<gateway-IP-number>:<netmask>:<client-hostname>:eth0:off
append initrd=initrd.img-2.6.31local ip=dhcp  root=/dev/nfs
nfsroot=/srv/fai/nfsroot boot=live  FAI_FLAGS=verbose,sshd,reboot
FAI_ACTION=install console=tty1 console=ttyS1


More information about the linux-fai mailing list