Remove NFS Storage

[sannsio]

Hi!

I have set up a NFS storage in my Proxmox Cluster. I now removed it, but the NFS share is still mounted on all the nodes:

192.168.1.235:/srv/nfs on /mnt/pve/NFSTest type nfs4 (rw,relatime,vers=4.2,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,clientaddr=192.168.1.80,local_lock=none,addr=192.168.1.235)

Will it be umounted automatically at some point of time?

Thanks a lot,
Sandra

 

[dylanw]

Hey,
What steps did you take to remove it?

 

[sannsio]

Hey,
What steps did you take to remove it?

I removed the storage in the the Storage section within the UI by clicking "Remove".
The NFS storage has been removed from the fiel /etc/pve/storage.cfg immediately.

 

[dylanw]

In this case, this is the intended behavior. Both the pvesm remove command line tool and the API only remove the configuration from Proxmox VE. It does not affect the underlying data or mount point, just PVE's awareness of it. To properly remove the mount, you would need to remove the entry from your /etc/fstab.

 

[sannsio]

In this case, this is the intended behavior. Both the pvesm remove command line tool and the API only remove the configuration from Proxmox VE. It does not affect the underlying data or mount point, just PVE's awareness of it. To properly remove the mount, you would need to remove the entry from your /etc/fstab.

Ok, thanks. Unfortunately there is no entry in my fstab for any mount:

root@pve1:/etc/pve# cat /etc/fstab # <file system> <mount point> <type> <options> <dump> <pass> /dev/pve/root / ext4 errors=remount-ro 0 1 /dev/pve/swap none swap sw 0 0 proc /proc proc defaults 0 0

Can I just umount the NFS mount manually? And do I have to do it on every host?

 

[dylanw]

Can I just umount the NFS mount manually? And do I have to do it on every host?

Yes you can unmount it manually, and this will have to be done on every host.

 

[fortechitsolutions]

Hi Forum, Thread small update in case anyone else hits this topic. I just bumped into this / came here via google search.

Context in my case:

- 3 node proxmox with mostly local storage 'share nothing' storage pools
- one NFS storage pool existed
- due to maintenance issues with hosting provider, the NFS storage pool server had to move to a new IP address. So NFS config needed adjust.
- I disabled the NFS storage pool in proxmox (ie, GUI > storage > NFS thing > Un-ticked the "enabled" box)
- then service provider migrated my NFS server, and gave it NEW IP address and then I tried to re-establish it as functional NFS storage in Proxmox.
- it seems no way to change an IP of an NFS storage in proxmox, so the path is to delete and re-create (I think).

the drama was that once I deleted the NFS storage pool config in proxmox. I was not able to add it back in with the new IP. giving me a timeout.

looking on the CLI, the output from 'mount' command still showed me the old mount was present
and output of "Dmesg" command tells me clearly lots of NFS TIMEOUT grumpy errors. For the no-longer-existing IP address of the NFS storage I tried to disable/remove.

note I have nothing in /etc/fstab related to this NFS mount. It was managed purely via GUI / proxmox storage UI.

Next step I tried to do a 'force' unmount and that tells me simply "Nope, device is busy, get bent!".

as per

umount -f /mnt/pve/nfs-misc



Then Finally I tried a "LAZY" unmount, and it seems the soft touch approach is better here.

umount -l /mnt/pve/nfs-misc

did not return any error, the NFS mount no longer is shown in output of "mount" command
and now bingo happy days I am proceed with adding the NFS storage back in / with the same name as before / but with a new IP address for the NFS server. And things 'just work fine'.

and of course I had to do the lazy-unmount thing on all 3 notes of proxmox cluster to make them all equally happy with the situation.

hope this info maybe of help to someone / or even just me in the future if I forget the solution and find it again via google in a year or two


Tim

 

[zha93]

I know this is an old thread... I'm rather new to proxmox and just discovered this "intended behavior" myself. Would someone please explain it to me? If I add a remote volume via the gui it mounts and works great. However removing it becomes a cli operation on every single node? I'm trying to understand the thought process as to why I need to manually unmount everything.

Thank you-

 

[Neobin]

For reference:
https://bugzilla.proxmox.com/show_bug.cgi?id=1016