NFS Shares hanging-on one of our PVE Hosts

[Guillaume Soucy]

Hello,

One of our Proxmox hosts is not able to connect to NFS shares anymore. Which process on that host need to be killed to for NFS shares to be reconnected?

I’d reboot the NFS server which is TrueNAS but, it didn’t fix the issue. I want to avoid rebooting the PVE host as it is a production server and this will cause downtime.

I’d tried to restart nfs-common by doing service nfs-common restart and I also try to kill [NFSv4 callback] without any success neither.

I’d also search the forums and search engines, found some information but wasn’t really helpful.

All our other PVE hosts are currently connected to the NFS server, the NFS server was under a heavy load last night as all our PVE hosts was sending backup it it at the same time.

Will review our backups strategy in a way they’re not stating at the same time. I just need to reconnect the disconnected PVE host to the NFS shares.

Thank-you,

Guillaume

 

[Guillaume Soucy]

Any idea? Please?

 

[Guillaume Soucy]

Need help please. Thank-you

 

[Guillaume Soucy]

I've got question marks at my NFS shares:

Last time I had to reboot to make it work. This is a not good, this is a production system, this is not normal to have to reboot to make NFS working again. It is stalled again, I will need to reboot again!

The NFS server is TrueNAS, I'd set the amount of servers to 256 which is the maximum amount possible and the issue still happening.

Need still help here.

Regards,

Guillaume

 

[Guillaume Soucy]

Hello,

I still having the same issue.

Need help please.

 

[corsinvest]

Hi Guillaume,
you could try
# ps -ef | grep vzdump
# kill -9 <processId>
to kill backup processes.
Task viewer eventually still shows killed backup jobs, but it's just a matter of gui, you fix it in a second time with reboot.

For unresponding NFS storage try to unmount and remount them, double checking your NFS server.

Best regards.

Alessandro

 

[Guillaume Soucy]

Hi Guillaume,
you could try
# ps -ef | grep vzdump
# kill -9 <processId>
to kill backup processes.
Task viewer eventually still shows killed backup jobs, but it's just a matter of gui, you fix it in a second time with reboot.

For unresponding NFS storage try to unmount and remount them, double checking your NFS server.

Best regards.

Alessandro

Hi,

Okay, I will try that next time it hang. However, it didn't happen again since I temporary disabled the backup task.

I will enable it again as of now and monitoring in the morning how it goes.

This is the 7th host, meaning its backups tasks only starts at 7AM.

I would like to keep tasks in this order in order to always have the same topology to base on when diagnosing.

Will keep updated.

Thank-you!

Guillaume

 

[Guillaume Soucy]

Hello,

It hangs again today so I replace the NIC. I suspect that I've got a bad NIC.

Will keep updated.

Guillaume

 

[Guillaume Soucy]

Still good. The previous NIC was probably bad.