PVE backup to NFS needs improvements !

[Petr Svacina]

Hi,

I have got big issues with backup task to NFS share.

The scenario:

You have mounted NFS share for the NFS backup task provided by vzdump. Everything works fine, until the NFS share is lost. ....

When the NFS share is starting to be not reachable during the backup, then the running backup task can not stop or be stopped, and the vm remain locked up.

At this point, I can not kill/stop correctly backup task, I can not stop the vm (even qm unlock ID is not working or qm stop ID). Is there any solution, how to destroy broken backup task ?

 

[dietmar]

This is standard NFS behavior. Try to use 'soft' flag when mounting NFS, or better - use a reliable NFS server.

 

[Petr Svacina]

This is standard NFS behavior. Try to use 'soft' flag when mounting NFS, or better - use a reliable NFS server.

Ok. I will try soft option.
"reliable NFS server" - every NFS server can be failed. But it should not block the PVE to be operational when backup is not longer possible, or if it is out of the control.

 

[dietmar]

"reliable NFS server" - every NFS server can be failed. But it should not block the PVE to be operational when backup is not longer possible, or if it is out of the control.

Again, this is defined by NFS - the idea is to wait until the server is back. The other option is to use 'soft' mount ...

see 'man nfs' for details.

 

[Petr Svacina]

Ok. Thanks for the quick reply.