copycat
copycat copied to clipboard
Use unique node identifiers
Currently, Copycat uses IP addresses as node identifiers. This can cause problems when migrating a node's data across IP addresses. A node's IP address should be dynamically updated when it joins/rejoins the cluster, and each node should instead use a unique string/numeric identifier instead.
👍 on this. Certainly makes sense to have some UID per instance.
I am trying to put some of cluster applications leveraging copycat cluster on kubernetes, and only using ip addresses cause difficulties since ip can always changes. Is there any "planned" schedule to implement this?