up vote 0 down vote favorite
share [g+] share [fb]

I have two machine, an NFS server (RHEL) and a client (Debian). The server has NFS set up, exporting a particular directory:

server:~$ sudo /usr/sbin/rpcinfo -p localhost
program vers proto   port
100000    2   tcp    111  portmapper
100000    2   udp    111  portmapper
100024    1   udp    910  status
100024    1   tcp    913  status
100021    1   udp  53391  nlockmgr
100021    3   udp  53391  nlockmgr
100021    4   udp  53391  nlockmgr
100021    1   tcp  32774  nlockmgr
100021    3   tcp  32774  nlockmgr
100021    4   tcp  32774  nlockmgr
100007    2   udp    830  ypbind
100007    1   udp    830  ypbind
100007    2   tcp    833  ypbind
100007    1   tcp    833  ypbind
100011    1   udp    999  rquotad
100011    2   udp    999  rquotad
100011    1   tcp   1002  rquotad
100011    2   tcp   1002  rquotad
100003    2   udp   2049  nfs
100003    3   udp   2049  nfs
100003    4   udp   2049  nfs
100003    2   tcp   2049  nfs
100003    3   tcp   2049  nfs
100003    4   tcp   2049  nfs
100005    1   udp   1013  mountd
100005    1   tcp   1016  mountd
100005    2   udp   1013  mountd
100005    2   tcp   1016  mountd
100005    3   udp   1013  mountd
100005    3   tcp   1016  mountd

server$ cat /etc/exports
/dir      *.my.domain.com(ro) 

client$ grep dir /etc/fstab
server.my.domain.com:/dir   /dir      nfs tcp,soft,bg,noauto,ro 0 0

All seems well, but when I try to mount, I see the following:

client$ sudo mount /dir
mount.nfs: access denied by server while mounting server.my.domain.com:/dir

And on the server I see:

server$ tail /var/log/messages
Mar 15 13:46:23 server mountd[413]: authenticated mount request from client.my.domain.com:723 for /dir (/dir)

What am I missing here? How should I be debugging this?

link|improve this question
What user owns /dir on server, and what user are you on client? – Bill Weiss Mar 15 '10 at 21:04
Two different users, but the server dir is set to 755 - readable by anyone. And since the export is RO, shouldn't that be sufficient? – zigdon Mar 15 '10 at 21:15
feedback

3 Answers

up vote 0 down vote

Is it broken from all machines or just one? Is the nfsd psuedo filesystem mounted on /proc/fs/nfsd on the server?

link|improve this answer
Server has the proc/fs/nfsd, but it's empty. It does show the exports in /proc/fs/nfs/exports though. Failing to mount from multiple machines. A network trace shows the server responding to the mount request with a ERR_ACCESS flag, makes me think it's something server side, not client. – zigdon Mar 15 '10 at 23:03
1  
If /proc/fs/nfsd is empty on server, it means the nfsd filesystem isn't mounted - it should have a few files in it which the NFS userspace uses to talk to the kernel server. Running mount -t nfsd none /proc/fs/nfsd should fix it. When it's not mounted you get this exact error from clients (been there, done that) – James Mar 16 '10 at 8:55
feedback
up vote 0 down vote

Unfortunately, rebooting the server solved the issue, so I still don't know why it was happening, but it's not anymore.

link|improve this answer
feedback
up vote 0 down vote

I've seen this if your /etc/hosts.allow and /etc/hosts.deny are not correct; check those files for a line with portmap in it and either comment it out (unsecure if you're not behind a firewall) or set the line on the client/server to be your specific subnet.

So for instance, in /etc/hosts.allow:

portmap: 192.168.0.0/16

...and comment out whatever is in /etc/hosts.deny to make only hosts.allow active. NFS uses tcpwrappers and these files to control access along with what's in /etc/exports.

link|improve this answer
feedback

Your Answer

 
or
required, but never shown
discard

Not the answer you're looking for? Browse other questions tagged or ask your own question.