[solved] CT stopped by nfs stale error, restart only by CLI?

[m.ardito]

Hi, now I soved this, but maube it's worth for someone to know:

I had a strange nfs stale error on my pve cluster (I had to umount -f -l the nfs share from gui and then it was accessible again).

then, I noticed that a CT was stopped, while it should have been running. Starting it from gui didn't work, failing with

TASK ERROR: mkdir /var/lib/vz/root/106: File exists at /usr/share/perl5/PVE/API2/OpenVZ.pm line 1100

and from cli I got

# ls -lah /var/lib/vz/root/
ls: cannot access /var/lib/vz/root/106: Stale NFS file handle
total 16K
drwxr-xr-x 5 root root 4.0K Nov 21 14:29 .
drwxr-xr-x 9 root root 4.0K Nov 19 12:09 ..
d????????? ? ?    ?       ?            ? 106
drwxr-xr-x 2 root root 4.0K Nov  8 17:18 107
drwxr-xr-x 2 root root 4.0K Sep 25 17:22 108

after some guessing I tried

~# vzctl start 106
Starting container ...
stat(/var/lib/vz/root/106): Stale NFS file handle
stat(/var/lib/vz/root/106): Stale NFS file handle
realpath(/var/lib/vz/root/106) failed: Stale NFS file handle
Container is unmounted
Container is mounted
Setting CPU units: 1000
Setting CPUs: 2
Configure veth devices: veth106.0
Adding interface veth106.0 to bridge vmbr0 on CT0 for CT106
Container start in progress...

and now it's running fine.

so, it happens that from web gui something is not able to do what cli does...

I don't know if this coudl be a bug or not. Worth to know, anyway.

Marco