I am trying to execute live (online) migration of an OpenVZ-container from one PVE cluster node onto another. Sometimes it does work, sometimes does not. I have noticed, that if container acts as a NFS-client (NFS is permitted for this container), then migration fails because of impossibility to suspend the machine.
Another such situation originates when container uses any folders of the host machine (zero container), mounted into the container using "mount --bind" or "mount --rbind" command.
Is this principial obstacles, or can I do something with them?
Another such situation originates when container uses any folders of the host machine (zero container), mounted into the container using "mount --bind" or "mount --rbind" command.
Is this principial obstacles, or can I do something with them?
