Backup destination disappears from the configuration file, breaking scheduled backups in OpsCenter
Scheduled backups stop running, throw "Destination missing" errors
"Cannot connect to cluster" error page appears in the UI
A cluster configuration file (located in /etc/opscenter/clusters/) used by the OpsCenter daemon is updated via a configuration management system (e.g. Chef, Puppet, Ansible), an automation script or a one-off user action on the machine where the OpsCenter daemon is hosted.
The OpsCenter daemon updates cluster configuration files during normal operation to keep track of cluster configurations. These are plain text, human-readable files. OpsCenter writes things like backup destinations and cluster connection information to these files when the relevant information is changed via the OpsCenter UI or API.
Having on-machine processes, such as configuration management jobs or user actions, overwrite these files can lead to accidental reverts or even omission of information written to these file by OpsCenter process.
To ensure this problem doesn't arise, we recommend not having these files managed by configuration management tools such as Chef or Puppet and defer to OpsCenter to maintain these files.
There is no immediate solution available that addresses this issue directly. Longer term we'd like to ensure configuration files are not written to by the OpsCenter process. There is no set timeline for this however. See OPSC-11667 for progress on this issue.