[SOLVED] NFS error

[daruom13]

s
Hello,

I deleted a storage from the web interface.
Since then, on one of the servers in the cluster, I have an error every second :

Nov 09 10:29:44 pve04 kernel: nfs: server ftpback-rbx6***** not responding, timed out

The hypervisor, VMs and storage appear with a "?".
But I still have access to their configuration.



In storage.cfg, it does not appear.



Thank you.

 

[daruom13]

anyone have an idea?

 

[fiona]

Hi,
removing a storage from the configuration does just that and doesn't "clean up". I guess the NFS server is probably still mounted on the node. Use umount -f -l /mnt/pve/<storage name of NFS> to unmount it.

 

[daruom13]

Hi,
removing a storage from the configuration does just that and doesn't "clean up". I guess the NFS server is probably still mounted on the node. Use umount -f -l /mnt/pve/<storage name of NFS> to unmount it.

Hello and thank you for your answer,

I performed the command, but the folder still appears in / mnt / pve / NFS_NAME.
The error is also present.
Should we restart the hypervisor?

 

[fiona]

Hello and thank you for your answer,

I performed the command, but the folder still appears in / mnt / pve / NFS_NAME.
The error is also present.

Can you check with findmnt if it's still mounted or if there is another NFS?

Should we restart the hypervisor?

If nothing else helps.

 

[daruom13]

Can you check with findmnt if it's still mounted or if there is another NFS?


If nothing else helps.

In the "problematic" server, the findmnt command does not return anything in particular :


On the other hand, on the 3 other hypervisors, the storage seems to be present.
However, they were deleted through the web interface and do not appear :




As this is a production server, I cannot restart until this evening (French time).
In addition, it blocks the migration and backup of the VMs that are on this server.
Is there a particular risk that can intervene and block the restart (as the error speaks of kernel) ?

 

[fiona]

In addition, it blocks the migration and backup of the VMs that are on this server.

That's strange as the storage is not configured anymore. Could you share the exact error and task log? Please also share the output of pveversion -v.

Is there a particular risk that can intervene and block the restart (as the error speaks of kernel) ?

I wouldn't think so, it's just that the NFS client runs within the kernel.

 

[daruom13]

That's strange as the storage is not configured anymore. Could you share the exact error and task log? Please also share the output of pveversion -v.


I wouldn't think so, it's just that the NFS client runs within the kernel.

Hello @Fabian_E ,

sorry to answer that now.
with the command pveversion -v, I noticed a difference in version on the kernel.
After updating the second server, I was able to migrate the VMs.
I was then able to restart the problematic hypervisor safely and it worked again normally.

Thanks again.

 

[Zaman]

That's strange as the storage is not configured anymore. Could you share the exact error and task log? Please also share the output of pveversion -v.


I wouldn't think so, it's just that the NFS client runs within the kernel.

is possible to restart it (nfs) without reboot system?
I try umount and vzdump --stop but after a seconds mounted automatically

 

[fiona]

Hi,

is possible to restart it (nfs) without reboot system?
I try umount and vzdump --stop but after a seconds mounted automatically

yes, Proxmox VE will mount it when it activates the storage. You need to disable the storage with pvesm set <storage ID> --disable to avoid that. Then unmount, restart the NFS server and activate the storage again with pvesm set <storage ID> --delete disable.

 

[crc-error-79]

Hi,
removing a storage from the configuration does just that and doesn't "clean up". I guess the NFS server is probably still mounted on the node. Use umount -f -l /mnt/pve/<storage name of NFS> to unmount it.

thank you very much!