Has any wisdom been established by the community or Ansible proper on which parts of a “backups infrastructure”, if any, Ansible could/should have a part in?
Ansible deployments obviously maintain quite a lot of data for deployment and life-cycle management that could be very useful to backup scripts (hosts of a particular type that need to be backed up, where the applications that need to be backed up were installed to, what user has rights to the data, let alone credentials to get into the hosts in the first place). Should this data (which must be maintained anyway) just be ignored and reimplemented in parallel for backups? Maybe regularly pulled into some dependent copy stored in another backend, which doesn’t get updated directly? (I haven’t followed the feature at all so I don’t know how it works or stores things, but “cached facts” spring to mind)
Could you write playbooks instead of backup scripts?
It seems to me that the only downside of pursuing this line of thought is that, well… they’re the freaking backups! A certain amount of bravery seems implied.
Thanks,
Mark