Unable to connect to VM socket after network storage fails

My Proxmox cluster uses nfs storage. Sometimes nfs server fails and VM remounts disks as readonly.

Trying to avoid this by adding werror=stop,rerror=stop to vm.conf like this:
ide0: nfs:101/vm-101-disk-1.qcow2,format=qcow2,rerror=stop,werror=stop,size= 10G

After this when nfs fails, it tryes to stop (pause) VM, but it can't connect to VM socket?

Here is what it says in syslog every 10 seconds:
pvestatd: WARNING: storage 'nfs' is not online
pvestatd: status update time (9.073 seconds)
pvestatd: WARNING: unable to connect to VM 101 socket - timeout after 31 retries

Any ideas?