I am trying to mount an NFS filesystem hosted by Celerra on our Linux
system - OS: SLES11 SP1
I get the following error: “mount.nfs: mount to NFS server
‘x.x.x.x:/sapextract’ failed: RPC Error: Program not registered”
Following is the output while checking nfs, rpcbind and portmap
services:
/etc/init.d/nfs status
Checking for mounted nfs shares (from /etc/fstab):
unused
/etc/init.d/rpcbind status
Checking for service rpcbind
running
/etc/init.d/portmap status
-bash: /etc/init.d/portmap: No such file or directory
Is there any specific configuration for SLES11 SP1 for NFS client?
NFS client on SLES10 was pretty staright forward
Your error “RPC Error: Program not registered” means the client asked
the server for some service but the server side says that service is not
running. There is a small chance that SLES11 SP1 is looking for a
version of a service, or a transport (TCP vs UDP) that is different that
what SLES10 asked for by default, but it is also likely that some daemon
has failed on the Celerra side.
Have the client do: rpcinfo -p <ip_of_celerra>
and make sure that the list which comes back includes all 5 services
involved with NFS usage: portmapper, mountd, nfs, lock manager, status.
If one of these is missing (or if you get the same rpc error) then
something is wrong on the Celerra side.
Thanks a lot for the reply. We have other clients (SLES10 SP2) that
mount NFS filesystems from Celerra without any issues. The issue happens
only with this particular host running SLES11 SP1. Following is the
output of rpcinfo on the SLES11 client
rpcinfo -p x.x.x.x
program vers proto port service
100000 2 tcp 111 portmapper
100000 2 udp 111 portmapper
100024 1 udp 32768 status
100024 1 tcp 32768 status
706000 1 udp 956
707000 1 udp 32829
I feel some configuration/service particular to SLES11 is missing on
the client side.
Please advise
The output you gave from rpcinfo -p is either not for the Celerra NFS
server or the Celerra server has serious problems. That output looks
more like the client’s own information. What address did you put in as
the x.x.x.x? As stated in my original reply, that was to be the ip
address of the Celerra. You can run the command from the client, but
it is to point to the Celerra’s address. If that is really the
Celerra’s information, then it is not even successfully running an NFS
daemon, nor a mount daemon, nor NFS lock manager daemon. All of which
should be running if the Celerra is an NFS server. Indeed, your error
says, “mount.nfs: mount to NFS server ‘x.x.x.x:/sapextract’ failed: RPC
Error: Program not registered” and according to the rpcinfo list, that
is true: The mount daemon is not registered (running) on your target.
So maybe your x.x.x.x was not correct, or maybe there is a duplicate IP
address on the network so you’re not reaching the host you think you are
reaching.
Do your other “working” clients get more services showing up when they
do “rpcinfo -p x.x.x.x” with the same Celerra IP address? Are those
other clients on the same subnet as the non-working client, or a
different subnet?
But putting aside the concern of whether the IP address is correct and
whether services are registered correctly on the Celerra, and just
considering other things that might be influencing your ability to mount
on the client side:
If the Celerra has more than one IP address bound to it, try pointing
your client’s mount request to the other IP address.
Try adding the parameters “proto=tcp,mountproto=tcp” to your client
mount options.
LAN trace analysis could be very helpful, especially if you analyzed
simultaneous packet traces taken from sles 11 nfs cleint and celerra nfs
server. If might also be useful to trace one of the working system
examples for comparison.