The hardware node is a part of Parallels Cloud Storage.
After changing the hostname of the node, all containers disappeared from it.
By design, the relation with the node, hosting containers on shared Cloud Storage, is provided by
.owner file inside container's private area. The file contains the node's hostname.
If the hostname changes, the container will lose its binding to the node and disappear from it completely.
To prevent the problem and avoid downtime of containers, it is necessary to migrate all containers from the hardware node to other nodes in online mode, change the hostname and migrate them back.
To fix the problem, once it already happened, register the containers again with
# prlctl register /vz/private/CTID --force # vzctl start CTID