On Thu, 2006-02-23 at 16:39 +0100, Steffen Winterfeldt wrote:
On Thu, 23 Feb 2006, Roger Oberholtzer wrote:
If that does not do, maybe you want NFS via tcp, boot with nfs.tcp=1 for that.
The kernel complains that there is no option called 'nfs.tcp=1', so it was ignored.
The option is for linuxrc. It normally uses v2 via udp. The above changes that to v3 and tcp.
Very interesting here:
Server: 192.30.105.201 (no ip) Looking for a network server... Trying to activate eth0 Setting up localhost...done eth0 activated Starting portmap OK, going to mount 0.0.0.0:/vol1/distro/OpenSUSE10.0/suse10.0:
...
if eth0 down
which obviously fails. Why 0.0.0.0? Just a few lines back it printed 192.30.105.201, which is the server. I do not know what the 'no ip' means.
Yes, sorry. The log is wrong with 0.0.0.0 at this point. Actually the server address is fine.
I tried to mount by hand, but the device was brought down (last line on console 3). There is no ifconfig command, so I was unsure how to bring up the device.
'ps' shows you whether dhcpcd is running. If not, just start it.
dhcpcd is not running, which makes sense as I used hostip= and netmask= on the command line. 'dhcpcd eth0' hangs. I see the request on the server, to which the server replies just like it did when the same client did a PXE boot some seconds earlier. But, unlike that time, the client does not respond to the dhcp address offered by the server.
BTW, in 10.1 some network tools are available for debugging at this point. You can pick up a beta mini boot-CD from opensuse.org if you want to try it. (You will need to boot with rootimage=/boot/root if you are using it with a 10.0 setup, but apart from that it will work fine.)
I'm not in the position to use 10.1 for this. I need to set up a production install of 10.0. -- Roger Oberholtzer OPQ Systems AB Ramböll Sverige AB Kapellgränd 7 P.O. Box 4205 SE-102 65 Stockholm, Sweden Tel: Int +46 8-615 60 53 Fax: Int +46 8-31 42 23