ntfs problem

Nigel Wade nmw at ion.le.ac.uk
Wed Mar 9 09:24:37 UTC 2005


Ankush Grover wrote:
> Hey friends,.
> 
> 
> I am having one 2 debian pcs and 1 FC3 pc in my office.I want to share
> some files from debian pcs so that users on FC3 can access it.I  can
> see the files by giving command
> 
> showmount -e 192.168.1.56. 
> 
> But whenever i try to mount that share it gives error
> 
> mount -t nfs 192.168.1.56:/ctn  /ctn
> 
> the /etc/exports file on one of the debian :
> 
> 
> /home/agrover           192.168.1.55(rw,sync,no_wdelay,squash uids-0-99)
> /ctn                          *(rw,async,no_root_squash)
> /var/log                      192.168.1.55(ro,all_squash)
> /raju                          192.168.1.55(rw,squash uids=0-99)
> /home/mverma            192.168.1.55(rw,squash uids=0-99)
> /home/rajat             192.168.1.55(rw,squash uids=0-99)
> /home/raju              192.168.1.55(rw,squash uids=0-99)
> 
> 
> 
> the error i get is 
> 
>  mount to NFS server '192.168.1.56' failed: server is down.The nfs
> ,portmap and rpc services are running on debain as i can see the
> process id with grep command.
> 
> Even in debian i have given /etc/hosts.allow 
> ALL:ALL
> 
> and also i have stopped iptables service in FC3 and SELINUX is disabled.
> Can anybody tell me what wrong i am doing.
> 
> Thanks & Regards
> 
> Ankush
> 

You will get the "server is down" error if the client can't find the correct 
NFS service from the server. This can be due to a firewall blocking 
portmapper, the NFS server not having registered itself with portmapper, or 
not supporting the version of NFS which the client is trying to mount.

To ensure that the client can see the server portmapper, and that the 
correct NFS version is registered run the following command on the client:

# rpcinfo -p <server>
where <server> is the hostname or IP of the server.

You should see various RPC services, e.g:

# rpcinfo -p some.host
    program vers proto   port
     100000    2   tcp    111  portmapper
     100000    2   udp    111  portmapper
     100024    1   udp  32768  status
     100024    1   tcp  32768  status
     391002    2   tcp  32769  sgi_fam
     100003    2   udp   2049  nfs
     100021    1   udp  32770  nlockmgr
     100021    3   udp  32770  nlockmgr
     100011    1   udp    617  rquotad
     100011    2   udp    617  rquotad
     100011    1   tcp    620  rquotad
     100011    2   tcp    620  rquotad
     100005    1   udp  32773  mountd
     100005    1   tcp  45994  mountd
     100005    2   udp  32773  mountd
     100005    2   tcp  45994  mountd

You need to see nfs and mountd.

Here, the server has only registered NFS vers. 2. To mount a filesystem from 
this server needs a '-o vers=2' option on the mount command.

In FC2 there appears to be a bug in portmap/nfs. Looking at the network 
traffic when I attempt to mount the filesystem with default args I can see 
the client first request the port for NFS V3 TCP. This isn't registered so 
portmap returns 0. The client next requests NFS V3 UDP, and portmap returns 
2049 (I think this is wrong, V3 isn't registered). The client then attempts 
a V3 UDP mount request on 2049 which fails and the client reports "server is 
down". This might be the problem you are seeing.

-- 
Nigel Wade, System Administrator, Space Plasma Physics Group,
             University of Leicester, Leicester, LE1 7RH, UK
E-mail :    nmw at ion.le.ac.uk
Phone :     +44 (0)116 2523548, Fax : +44 (0)116 2523555




More information about the fedora-list mailing list