Getting RPC error while mounting

[Harsh_Sharma]

Hi,
I am trying to mount a filesystem from Solaris machine to HP-UX machine and see the following error:
# mount -F nfs jerry:/blr /space/blr
nfs mount: get_fh: jerry:: RPC: Program not registered
nfs mount: get_fh: jerry:: RPC: Program not registered
nfs mount: retry: retrying(1) for: /space/blr after 5 seconds
nfs mount: retry: giving up on: /space/blr

The nfs server and client are up and running. Also rpc.mountd is also up and running.

[Peter Godron]

Hi,
and welcome to ther forums !

Please see earlier thread:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=243112

Please also read:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33 on how to reward any useful answers given to your questions.

[Steven E. Protter]

Shalom,

At startup.

nfs.core
nfs.server
nfs.client

All must be configured correctly to be a server and client. core must run before both.

You need to check the configuration on the Solaris box and insure no firewalls or switch filtering is blocking the traffic.

SEP

[Harsh_Sharma]

Hi Peter,
I did read the above thread before submitting the question. But i am still getting the error.
Thanks,
Harsh

[Peter Godron]

Hi,
thanks for confirming you had read the earlier thread.

NFS is pretty straight forward:
Server - network - client

You confirmed server and client are running.
Doubvle check the conf files are ok. This would then rule out server and client and leave network.

Can you check that the host jerry is known and accessable from the client.

[Harsh_Sharma]

Hi Steve,
I shut down the core, server and client and restarted them in order - core, server , client. But even then i am seeing the same prob.
Thanks
Harsh

[Harsh_Sharma]

Hi Peter/Steve,
Thanks for the quick resposes...

Peter,

I checked the conf file (/etc/rc.config.d/nfsconf) and checked the entries for NFS_CLIENT and NFS_SERVER and both are set to 1. Also START_MOUNTD=1 and AUTOFS=1 are set.

One thing that i did on the Solaris machine was to name the nfs domain as nmsnmi.com. I do not know how to name a nfs domain on HP. (Actually i had to do that in order to solve mount prob on one of my AIX machine)

Does this make any difference?
Thanks,
Harsh

[Dave Olker]

Hi Harsh,

It sounds like the HP-UX system is the NFS client, correct? What OS is each system running? From the HP-UX system issue the following commands:

# rpcinfo -p
# showmount -e

and post the output please.

Regards,

Dave

[Harsh_Sharma]

Hi Dave,
HP-UX is 11.23 and Solaris ver is 10.
The output of the cmds are:
# rpcinfo -p 10.31.76.117
program vers proto port service
100000 4 tcp 111 rpcbind
100000 3 tcp 111 rpcbind
100000 2 tcp 111 rpcbind
100000 4 udp 111 rpcbind
100000 3 udp 111 rpcbind
100000 2 udp 111 rpcbind
100024 1 udp 32772 status
100024 1 tcp 32771 status
100133 1 udp 32772
100133 1 tcp 32771
100021 1 udp 4045 nlockmgr
100021 2 udp 4045 nlockmgr
100021 3 udp 4045 nlockmgr
100021 4 udp 4045 nlockmgr
100021 1 tcp 4045 nlockmgr
100021 2 tcp 4045 nlockmgr
100021 3 tcp 4045 nlockmgr
100021 4 tcp 4045 nlockmgr
100001 2 udp 32773 rstatd
100001 3 udp 32773 rstatd
100001 4 udp 32773 rstatd
100002 2 tcp 32772 rusersd
100002 3 tcp 32772 rusersd
100002 2 udp 32774 rusersd
100002 3 udp 32774 rusersd
100011 1 udp 32775 rquotad
100083 1 tcp 32773 ttdbserver
100068 2 udp 32776 cmsd
100068 3 udp 32776 cmsd
100068 4 udp 32776 cmsd
100068 5 udp 32776 cmsd
300598 1 udp 32779
300598 1 tcp 32777
805306368 1 udp 32779
805306368 1 tcp 32777
100249 1 udp 32780
100249 1 tcp 32778
100068 2 tcp 35019 cmsd
100068 3 tcp 35019 cmsd
100068 4 tcp 35019 cmsd
100068 5 tcp 35019 cmsd
# showmount -e 10.31.76.117
showmount: 10.31.76.117: RPC: Program not registered
#

[Yang Qin_1]

Hi, can you verify /etc/rc.config.d/nfsconf? make sure
NFS_CLIENT=1
NFS_SERVER=1

Yang

[Harsh_Sharma]

Hi Yang,
The entries are 1 for both NFS_SERVER and NFS_CLIENT.
Thanks,
Harsh

[Dave Olker]

Hi Harsh,

The problem is the nfsd and mountd daemons are not running on the Solaris system. The way Solaris works is you have to share a filesystem in order for these daemons to start.

Have you shared any NFS filesystems on the Solaris system? What is the conents of the files:

/etc/dfs/dfstab
/etc/dfs/sharetab

If you type the command "share" on the Solaris system do you see any filesystems shared for NFS access?

Regards,

Dave

[Harsh_Sharma]

Hi Dave,
The contents of the two files are:
jerry-root-/blr % cat /etc/dfs/sharetab
/blr - nfs rw
/jainr - nfs rw
jerry-root-/blr % cat /etc/dfs/dfstab

# Place share(1M) commands here for automatic execution
# on entering init state 3.
#
# Issue the command 'svcadm enable network/nfs/server' to
# run the NFS daemon processes and the share commands, after adding
# the very first entry to this file.
#
# share [-F fstype] [ -o options] [-d ""] [resource]
# .e.g,
# share -F nfs -o rw=engineering -d "home dirs" /export/home2
share -F nfs -o rw /blr
share -F nfs -o rw /jainr
jerry-root-/blr %

Here /blr is the share that i want to mount on HP machine.

But the showmount cmd on Solaris results in;
jerry-root-/blr % showmount
showmount: jerry: RPC: Program not registered

Thanks,
Harsh

[Dave Olker · Accepted solution]

Hi Harsh,

For some reason the nfsd and mount daemons didn't get started at boot time on the Solaris system. They are supposed to be launched automatically when you issue the command "share -a", which is done during boot time.

Please try issuing the command "share -a" and then repeat the showmount command and see if the NFS server daemons were started.

Regards,

Dave

[Peter Godron]

Hi,
there seem to be a lot of similar NFS problems lately, although on different versions:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1078361

In your earlier post you confirmed everything was running.

[Harsh_Sharma]

Hi Dave,
I think you got it right. I just rebooted the system to get and the mountd daemon started and thus the problem was solved.

Thanks for all your support.

Thanks,
Harsh

[Harsh_Sharma]

Hi Peter,
I checked on HP machines but did not check on the Solaris machine. Initially i was working on AIX machine (for mounting the same FS of Solaris). I was able to mount the FS but during that time i think something happened and mountd stopped on Solaris box.
Thanks for all the support.
Thanks,
Harsh

[Dave Olker]

Hi Peter,

Yeah, I noticed that too. Seems like lots of people are reporting RPC errors trying to mount lately. I've helped out with several of them and they are not really related. Some of them turn out to be hostname resolution stuff, while others, like this one, are simple "daemons aren't running" stuff. Strange how many hit the forums all at once for different reasons.

This one, involving Solaris 10, could be caused by the fact that Solaris doesn't start nfsds and mountd at boot time unless there is an explicitly shared filesystem in the /etc/dfs/dfstab file at boot. In other words, if you don't have any filesystems configured for sharing at boot time the daemons won't get started.

This is a very different approach from HP-UX and other implementations where you typically configure whether you want these daemons started in some configuration file (like /etc/rc.config.d/nfsconf on HP-UX). If you tell us to run the daemons we start the daemons regardless of whether there are filesystems exported/shared or not.

The Solaris approach is "we'll only start these daemons if there's a potential for us to actually do some NFS server work, so if you don't share any filesystems there's no reason to start the daemons." The problem is many people get bit by this where they configure the filesystems in the /etc/dfs/dfstab file and then expect the daemons to just be always running. In the case of Solaris 10 you have to explicitly call "share -a" to get the daemons to start if they weren't started at boot time (because there were no filesystems shared at boot time). I've seen this bite many a SysAdmin, especially those that administer both HP-UX and Solaris.

Regards,

Dave