Forums

Home / Forums

You need to log in to create posts and topics. Login · Register

Can't get NFS exports to mount

Hey there,

This is a bit of an odd one for me. I have a cluster up and running on 2.7.1 latest ISO, I have CephFS running and CIFS is working correctly.

I have NFS configured, everything appears to be configured correctly. I can view the ganesha containers are up and running and serving over their IP's, and if I watch the Petasan.log it says that  NFS exports service is started and it has succeeded in acquiring the resource.

However, when I attempt to mount the export on a client machine, it just hangs. I attempted to use verbose mode to see if there is any reason, but there is no errors - it just simply times out.

I wondered if it might have something to do with the client itself, so to sanity check I attempted to mount the NFS export locally on one of the petasan nodes - and same thing. There is no indication of whats wrong in the logs as it does not give any errors, and it finishes simply with a timeout.

Have you ever come across this, and is there any other more verbose logs I might be able to access to see if there is something I am missing?

Thanks!

By the way - I am really enjoying Petasan! Some really cool choices.

can you ping the ip ? are you using a routed client ? are you using a custom gateway for NFS subnet ?

assuming you have an export named export1

on client use:
mkdir -p /mnt/export1
mount -t nfs -o nfsvers=4.1,proto=tcp <IP>:/export1 /mnt/export1/

replace <IP> with public ip or round robin dns name

Hi Admin thanks for the reply,

 

I can ping the IP.

I am using a routed client and a custom gateway for the NFS subnet, however the client can ping the NFS vIP's

That is the syntax I am using for the client mount attempt.

Here is what it says and just hangs

mount.nfs: timeout set for Wed Jan 27 09:30:34 2021

mount.nfs: trying text-based options 'vers=4.1,addr=192.168.10.45,clientaddr=192.168.20.4'

unfortunately we have an issue with custom routes with NFS, it is stable with CIFS  but not in NFS. Can you try with default route and use clients having ips on the NFS public subnet (ie non routed )

Thanks for the reply.

You are correct! This was the issue. Thank you for the info as I was wracking my brain trying to understand what was wrong.

Is there any ETA on a fix for this?