Bug fix: Remote restores should only be eligible on uninitialized clusters

[davissp14]

Problem
After a cluster is restored, the restore.lock is set on the primary to prevent the restore from re-triggering on reboot. However, the restore.lock is not set on the replicas so when they are rebooted the restore process will kick in and blow away the repmgr database...

If you hit this issue, you will see a log on your replica saying: Restoring from a hot standby., with the following logs:

2024-07-26T01:48:22Z app[8d9511aee26008] ord [info]postgres | 2024-07-26 01:48:22.876 UTC [14421] FATAL:  database "repmgr" does not exist
2024-07-26T01:48:22Z app[8d9511aee26008] ord [info]failed post-init: failed to register new standby: failed to register standby: exit status 1. Retrying...

Solution
The restore process is now being properly restricted to uninitialized clusters. If the cluster has already been initialized, meaning the cluster's primary has been established, the restore process will be ignored.