According to error message there is some VZFS (Virtuozzo filesystem) inconsistency inside a VE running on the node (which could happen, for example, due to unclean server shutdown). Usually in such error messages inode # and COW (copy-on-write) filename also reported, all COW files are usually located in /vz/private/VEID/cow/, so it should be possible to locate the problem VE using just for example this command:

# ls -l /vz/private/*/cow/NAME

where NAME is filename reported in the error message.

To fix the problem 'vzfsutil' utility should be executed against unmounted (and stopped) VE private area:

# vzfsutil --call=move --oall=do -t /vz/template /vz/private/VEID

where VEID is ID of problem VE shown by 'ls' command in previous section.