root-nfs can not start up

PhillipHuang · 12-17-2012, 03:07 AM

Hello all,

I have encounted "Unable to get nfsd port number from server" when start up a root-nfs system via tftp.

Machine A: centos5.4, dhcp/tftp server 172.16.30.253
Machine B: client which support PXE (dhcp_ip: 172.16.30.254)

First, I mount a HD on machine B and rebuild the kernel which can support root-nfs, and then copy the necessary directories and files to the NFS share "/mnt/c" on machine A. Here's the configuration file on Machine A,

/etc/dhcpd.conf

Code:

ddns-update-style interim;
ignore client-updates;
next-server 172.16.30.253;
filename "/pxelinux.0";

subnet 172.16.30.0 netmask 255.255.255.0 {

        option subnet-mask              255.255.255.0;
        option domain-name-servers      172.16.30.253;
        option time-offset              -18000; # Eastern Standard Time
        range dynamic-bootp 172.16.30.128 172.16.30.254;
        default-lease-time 21600;
        max-lease-time 43200;

}

/tftpboot/pxelinux.cfg/default

Code:

default Linux
prompt 0
label Linux
  kernel vmlinuz
  append init=/sbin/init root=/dev/nfs rw nfsroot=172.16.30.253:/mnt/c ip=dhcp vga=text

/etc/hosts.allow

Code:

in.tftp: 172.16.30.254

/etc/exports

Code:

/mnt/c	*(rw,sync,no_all_squash,no_root_squash)

service status,dhcpd and xinetd are both running.

Code:

# service nfs status
rpc.mountd (pid 2558) is running...
nfsd (pid 2555 2554 2553 2552 2551 2550 2549 2548) is running...
rpc.rquotad (pid 2526) is running...

Machine A and B are connnected to a cisco switch. when poweron B, it encounted:

Code:

Looking up port of RPC 100003/2 ON 172.16.30.253
rpcbind: server 172.16.30.253 not responding, time out
Root-NFS: Unable to get nfsd port number from server, using default

Looking up port of RPC 100005/1 ON 172.16.30.253
rpcbind: 172.16.30.253 not responding, time out
Root-NFS: Unable to get nfsd port number from server, using default

mount:server 172.16.30.253 not responding, time out
Root-NFS: Server returned -5 while mounting /nfsroot

I have googled many articles, but none can help me fix it. Would you please give me any suggestion?

Thanks in advanced,

Phillip

jason_not · 12-18-2012, 06:11 PM

Hello,

My first two thoughts are "is portmapper running on the nfs server" and "is there a firewall running on the nfs server"

A bit of history: rpc ports are not the same as "tcp" or "udp" ports. the rpc portmapper listens on tcp and/or udp port 111, incase you have trouble with the firewall.

on nfs version 3 and below:

On the nfs server, you should be able to type "rpcinfo -p" and get a list of the various rpc services running. If there is an error, then portmapper might not be running.

if this works, AND you have a separate linux system, you can run on the other linux system "rpcinfo -p <nfs server ip>" to see if the firewall is blocking access to the portmapper.

--jason

PhillipHuang · 01-25-2013, 12:12 AM

Thank you, Jason.

I did not build the correct driver for on board Intel Network-device. After built-in and recompiled, it was fixed.

jason_not · 01-25-2013, 12:18 AM

Hey, that's really nice to hear! thanks for posting!