Problem with nfs and initramfs-tools
Hello Kernel-Team,
I am using a Debian-based nfs-root as a rescue-system in our datacenter.
I am running into some problems with the initial ramdisks and nfsroot.
In 90% everythiong is OK, but 10% of our servers have a second nic. In
these cases the server will not boot, as he knows not how to mount the
nfs-root.
/usr/src/linux/Documentation/nfsroot.txt explains the parameters neded
and describes the error, which occours in my setup:
ip=<client-ip>:<server-ip>:<gw-ip>:<netmask>:<hostname>:<device>:<autoconf>
If this parameter is missing from the kernel command line, all fields
are assumed to be empty, and the defaults mentioned below apply. In
general this means that the kernel tries to configure everything using
autoconfiguration.
I am using the following commandline in my pxe-config:
append initrd=initrd_rescue.2 root=/dev/nfs
nfsroot=192.168.123.123:/export/root/nfsroot ip=::::::dhcp
The system gets an IP and I am able to ping the system, but there is an
ifinite loop trying to mount the rootfs. As far as I can see in the
output of the console, the rootpath is missing:
rootserver:192.168.123.123 rootpath:
filename: pxelinux.0
If I put the commandline to
append initrd=initrd_rescue.2 root=/dev/nfs
nfsroot=192.168.123.123:/export/root/nfsroot ip=:::::eth1:dhcp
evrything works, but as far as I understand it should be not necessary
to define the interface. It should wor automatically.
I have no idea how to track down the problem, I hope, you can give me a
hint.
--
Mit freundlichen Gruessen
Andre Timmermann
Nine Internet Solutions AG, nine.ch, Letzigraben 77, CH-8003 Zuerich
Tel +41 44 637 40 00 | Direkt +41 44 637 40 06 | Fax +41 44 481 16 43
Reply to: