Hello, I'm having difficulty getting a rebuilt (OS Re-installed) server to mount a previously mounted NFS share.

mount.nfs: trying text-based options 'vers=3,rsize=32768,wsize=32768,addr=x.x.x.99'
mount.nfs: prog 100003, trying vers=3, prot=6
mount.nfs: trying x.x.x.99 prog 100003 vers 3 prot TCP port 2049
mount.nfs: prog 100005, trying vers=3, prot=17
mount.nfs: trying x.x.x.99 prog 100005 vers 3 prot UDP port 57867
mount.nfs: mount(2): Stale file handle


All the other mounts work properly (roughly 9 others). We are using 2 different NIC types. One for data traffic (10gb network) and other for management (1gb network).

The NFS client has been rebooted, the NFS server has had the nfsd services restarted, previously tried exportfs -rf.

One oddity, in rmtab I'm showing the IP address of the 1gb network and the 10gb network for the client. Currently on the client, everything is set to connect to the 10GB interface, so I'm not certain why it is showing on the 1GB connection.

1GB Network:
[root@prdhostname ~]# cat /var/lib/nfs/rmtab | grep 154
x.x.x.154:/data01/share01:0x00000004

10GB Network:
[root@prdhostname ~]# cat /var/lib/nfs/rmtab | grep 120
x.x.x.120:/data02/share02:0x00000024
x.x.x.120:/data01/share01:0x00000015

Any advice / troubleshooting tips would be appreciated.

What is in your exportfs file? Did you restrict the lines there to the 10 GB network IP range or did you do it by hostname? If the latter is it possible you have the hostname being found by its 1 GB NIC IP (in /etc/hosts, DNS or NIS)?

Also you might want to check the NFS client server to see how it discovers the name of the NFS server (hosts, NFS or NIS) to be sure it is using the 10 GB name.

That is a good point. I'll look in to this and let you know what I find.

Thanks!

if I would like to command mount into a small box, there will be 2 machines connecting by a cable.

again, only 2 machines and my question was: which machine box is the server side?
the machine where I type the mount command OR the machine I successfully mounted into ?

please try command(s)
mount.nfs 192.168.1.1:/whichDirectory /media -o port=80 -v
mount.nfs4 192.168.1.1:/whichDirectory /media -o mountport=80 -v
many connection refused then Connection timed out

You really shouldn't piggyback on other threads. You should open new ones as you're more likely to get responses. Existing threads often are only looked at by the original owner and anyone who responded.

In answer to your question as it relates to NFS the "server" side is that which has the local filesystem that you exported to other systems. The "client" side is the one that mounts what you exported from the "server".

Vagary in the language: Server can also mean "system" or "host" so often one might talk about the "client server" when meaning "client system" or "client host". Server can also mean a system designed for multiple users as opposed to a workstation used by only one. In the NFS usage the meaning is that implied by "server, client" setup when speaking of applications/facilities wherein one side (the "server") is running something that the other side (the "client") uses. In Linux any "system" or "host" (and even most workstations) can be either the "server" or the "client" for applications/facilities.

It does sometimes occur that you would have one "system" or "host" be the "server" for one application/facility with another being the "client" of that but the other system being the "server" for another application/facility and the original system being the "client" for the other application/facility. e.g. You could have two systems both as NFS servers and as NFS clients where each exports a filesystem to the other.

thank you for replying.

to mount into a router box, should the fs type option -t

cifs ? mount.cifs ? the output was Unable to find suitable address

You need to open a new thread as I told you before.