View previous topic :: View next topic |
Author |
Message |
dusanc Apprentice
Joined: 19 Sep 2005 Posts: 248 Location: Serbia
|
Posted: Wed Apr 16, 2008 5:41 pm Post subject: [SOLVED] Diskless install, NFS root, kernel can't mount / |
|
|
Hi,
I'm trying to install diskless gentoo using getoo altinstall guide http://www.gentoo.org/doc/en/altinstall.xml
I've managed to start with pxegrub and it loads the kernel from server, but kernel stops when it tries to mount / with following (hand copied):
Code: | Looking up port of RPC: 100005/1 on 192.168.1.3
rpcbind: Server 192.168.1.3 not responding. Timed out |
And then it says it can't mount root fs.
Relevant info:
Code: | cat /etc/exports
/diskless/192.168.1.4 192.168.1.4(nohide,rw,sync,no_root_squash)
|
Code: | cat /diskless/192.168.1.4/boot/grub.lst
default 0
timeout 30
title=Diskless gentoo
root (nd)
kernel /192.168.1.4/boot/bzImage in=dhcp root=/dev/nfs nfsroot=192.168.1.3:/diskless/192.168.1.4
|
dhcp.conf
Code: | option option-150 code 150 = text;
option root-path "192.168.1.3:/diskless/192.168.1.4";
next-server 192.168.1.3;
option routers 192.168.1.1;
shared-network KRSHEVI {
subnet 192.168.1.0 netmask 255.255.255.0 {
range 192.168.1.250 192.168.1.254;
host krshina1 {
hardware ethernet 00:19:E0:0B:8A:6D;
fixed-address 192.168.1.4;
option option-150 "/192.168.1.4/boot/grub.lst";
filename "/192.168.1.4/boot/pxegrub"; |
Code: |
rpcinfo -p 192.168.1.3
program vers proto port
100000 2 tcp 111 portmapper
100000 2 udp 111 portmapper
100024 1 udp 46851 status
100024 1 tcp 44758 status
100005 1 udp 57728 mountd
100005 1 tcp 56867 mountd
100005 2 udp 57728 mountd
100005 2 tcp 56867 mountd
100005 3 udp 57728 mountd
100005 3 tcp 56867 mountd
100021 1 udp 51853 nlockmgr
100021 3 udp 51853 nlockmgr
100021 4 udp 51853 nlockmgr
100003 2 udp 2049 nfs
100003 3 udp 2049 nfs
100021 1 tcp 60350 nlockmgr
100021 3 tcp 60350 nlockmgr
100021 4 tcp 60350 nlockmgr
100003 2 tcp 2049 nfs
100003 3 tcp 2049 nfs
|
Ofcourse client kernel (v2.6.24-gentoo-r4) is compiled with NFS and root over NFS support and following:
Code: |
[*] TCP/IP networking
│ │ [ ] IP: multicasting
│ │ [ ] IP: advanced router
│ │ [*] IP: kernel level autoconfiguration
│ │ [*] IP: DHCP support
│ │ [*] IP: BOOTP support
|
This is what I get in servers /var/log/messages:
Code: | Apr 16 19:58:11 krshina3 dhcpd: DHCPRELEASE from 00:19:e0:0b:8a:6d specified requested-address.
Apr 16 19:58:11 krshina3 dhcpd: DHCPRELEASE of 192.168.1.4 from 00:19:e0:0b:8a:6d via br0 (not found)
Apr 16 19:58:37 krshina3 dhcpd: DHCPDISCOVER from 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:37 krshina3 dhcpd: DHCPOFFER on 192.168.1.4 to 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:40 krshina3 dhcpd: DHCPREQUEST for 192.168.1.4 (192.168.1.3) from 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:40 krshina3 dhcpd: DHCPACK on 192.168.1.4 to 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:40 krshina3 dhcpd: DHCPDISCOVER from 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:40 krshina3 dhcpd: DHCPOFFER on 192.168.1.4 to 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:40 krshina3 dhcpd: DHCPREQUEST for 192.168.1.4 (192.168.1.3) from 00:19:e0:0b:8a:6d via br0
Apr 16 19:58:40 krshina3 dhcpd: DHCPACK on 192.168.1.4 to 00:19:e0:0b:8a:6d via br0
|
And just to mention I used LTSP and it worked with NFS.
Thanks in advance
Dushan
Last edited by dusanc on Wed Apr 23, 2008 8:45 pm; edited 1 time in total |
|
Back to top |
|
|
Januszzz Guru
Joined: 04 Feb 2006 Posts: 367 Location: Opole, Poland
|
Posted: Thu Apr 17, 2008 6:54 am Post subject: |
|
|
Hi again,
can you mount this NFS share from another machine or from a server? adjust /etc/exports to permit others to mount for the moment.
Is your portmap running and netfilter isn't blocking any connections?
I remember having somewhat similar problem months ago.... here it is:
https://forums.gentoo.org/viewtopic-t-566068-highlight-.html
EDIT: Then I assumed that it was a problem with nfs and thus I changed directories, as drowning catch a knife too. It helped surprisingly |
|
Back to top |
|
|
dusanc Apprentice
Joined: 19 Sep 2005 Posts: 248 Location: Serbia
|
Posted: Thu Apr 17, 2008 5:00 pm Post subject: |
|
|
I can mount the root NFS share from the server:
Code: | krshina3 dusan # mount -t nfs 192.168.1.3:/diskless/192.168.1.4 /mnt/test/
krshina3 dusan # ls /mnt/test/
bin boot dev etc home lib mnt opt proc root sbin sys tmp usr var
|
I don't use netfilter on this server. I tried to change the directory, but still nothing.
How can I confirm if the client is actually sending those requests, and does the server recieves them?
Thanks
Dushan |
|
Back to top |
|
|
buzzin Apprentice
Joined: 17 Oct 2003 Posts: 264 Location: St. Albans, UK.
|
Posted: Sat Apr 19, 2008 2:48 pm Post subject: |
|
|
dusanc wrote: | How can I confirm if the client is actually sending those requests, and does the server recieves them? |
You could try running tcpdump on the server and see if 192.168.1.3 is making a tcp connection at all. |
|
Back to top |
|
|
dusanc Apprentice
Joined: 19 Sep 2005 Posts: 248 Location: Serbia
|
Posted: Sat Apr 19, 2008 3:59 pm Post subject: |
|
|
This is what I get
Code: | tcpdump -vvvi br0 host 192.168.1.4
17:55:56.054507 IP (tos 0x0, ttl 64, id 1987, offset 0, flags [DF], proto UDP (17), length 1464) krshina3.gekom.com.30094 > krshina1.gekom.com.2002: UDP, length 1436
17:55:56.054870 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.054885 IP (tos 0x0, ttl 64, id 1988, offset 0, flags [DF], proto UDP (17), length 1464) krshina3.gekom.com.30094 > krshina1.gekom.com.2002: UDP, length 1436
17:55:56.055245 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.055263 IP (tos 0x0, ttl 64, id 1989, offset 0, flags [DF], proto UDP (17), length 1464) krshina3.gekom.com.30094 > krshina1.gekom.com.2002: UDP, length 1436
17:55:56.055620 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.055636 IP (tos 0x0, ttl 64, id 1990, offset 0, flags [DF], proto UDP (17), length 1464) krshina3.gekom.com.30094 > krshina1.gekom.com.2002: UDP, length 1436
17:55:56.055993 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.056350 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.056718 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.057100 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
17:55:56.057115 IP (tos 0x0, ttl 64, id 1994, offset 0, flags [DF], proto UDP (17), length 176) krshina3.gekom.com.30094 > krshina1.gekom.com.2002: UDP, length 148
17:55:56.057210 IP (tos 0x0, ttl 60, id 0, offset 0, flags [none], proto UDP (17), length 32) krshina1.gekom.com.2002 > krshina3.gekom.com.30094: [udp sum ok] UDP, length 4
739 packets captured
1608 packets received by filter
869 packets dropped by kernel
|
192.168.1.3 = krshina3 = server
192.168.1.4 = krshina1 = client (don't ask about why I choose these IP adresses )
What now? |
|
Back to top |
|
|
dusanc Apprentice
Joined: 19 Sep 2005 Posts: 248 Location: Serbia
|
Posted: Sat Apr 19, 2008 4:39 pm Post subject: |
|
|
When I select only dhcp and not bootp under clients kernel conf, that kernel panics on boot.
Is it possible that clients kernel only tries to connect to bootp service and not to dhcp? |
|
Back to top |
|
|
dusanc Apprentice
Joined: 19 Sep 2005 Posts: 248 Location: Serbia
|
Posted: Wed Apr 23, 2008 8:45 pm Post subject: |
|
|
I solved it
Kernel level autoconfiguration didn't work so I just put in parameters that should be given by dhcp in the grub.lst as kernel command line and shut off the autoconfiguration like this:
Code: | kernel /4/boot/bzImage ip=192.168.1.4:192.168.1.3:off root=/dev/nfs nfsroot=192.168.1.3:/diskless/4
|
After that I had to remember that tmpfs has to be configured in kernel or mounting /dev for UDEV wouldn't work
Shouldn't this be emphasised in handbook? |
|
Back to top |
|
|
SeaTiger l33t
Joined: 22 Nov 2007 Posts: 603 Location: Toronto, Ontario, Canada
|
Posted: Thu Apr 24, 2008 4:59 am Post subject: |
|
|
I think I find your original error in pxe config file:
Code: | cat /diskless/192.168.1.4/boot/grub.lst
default 0
timeout 30
title=Diskless gentoo
root (nd)
kernel /192.168.1.4/boot/bzImage in=dhcp root=/dev/nfs nfsroot=192.168.1.3:/diskless/192.168.1.4 |
The last line, "in=dhcp" should be "ip=dhcp". |
|
Back to top |
|
|
dusanc Apprentice
Joined: 19 Sep 2005 Posts: 248 Location: Serbia
|
Posted: Thu Apr 24, 2008 5:37 am Post subject: |
|
|
LOL
Thanks.
I saw that one too but after 8 days. The thing is it doesn't work when I set it to ip=dhcp, so I just put the fixed parameters and voila, problem solved |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|