I can see that copr fe has some progress on making the db backed-up by rdiff-backup: https://fedora-infra-docs.readthedocs.io/en/latest/sysadmin-guide/sops/rdiff-backup.html
But it doesn't seem to be finished, namely we are not in appropriate ansible inventory group. Can I please add it there, and let the DB backup automatically? ATM, it is less than ~100MB daily.
Ad, Restoring from backups from SOP:
Restoring from backups
rdiff backup keeps a copy of the most recent version of files on disk, so if you wish to restore the last backup copy, simply rsync from backup03. If you wish an older incremental, see rdiff-backup man page for how to specify the exact time.
I suppose I can not do this myself, right? So I'd have to fill a ticket asking you to provide me the backup.
Ok, I was told today that we'd like to also backup copr keygen database, which is ~110MB daily ATM.
Metadata Update from @kevin: - Issue priority set to: Waiting on Assignee (was: Needs Review) - Issue tagged with: high-gain, low-trouble
ok, I sat down to do this, but found I have some questions for you:
On copr-fe you only want the db dump in /backups backed up? nothing else ?
On copr-keygen, where is the database? Can it be dumped to /backups ? Once it is, you only want that backed up, nothing else?
We don't use rdiff-backup to backup db dumps anymore as it's very inefficent at doing that, so we just copy each days database dump to the backup server. So, we have a daily database dump for the last N days (where N is very large). We do use rdiff-backup to backup other directories (by default /etc and /home on machines we keep backups of, but we can add other directories).
Yes, everything else can be spawned from scratch by playbook.
On keygen, there's no database, just a set of generated gpg keys. That is dumped to host_backup_targets: ['/backup/'] ATM. I can change the ldayout if that was needed.
host_backup_targets: ['/backup/']
TBH, I don't think it matters "how" it is done. So I'm glad we can setup the backups!
What would be nice though to be able to check from time to time that the backup is working (have some access to the backed-up tarballs). If that isn't possible, we can check only the /backup or /backups directories only and trust you that everything is fine.
ok. I should have this setup. We can see after tomorrow's backup run.
If you like you could put in a ticket from time to time to restore a backup from some past time, we could put it somewhere for you and you could check it?
Metadata Update from @kevin: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
ack, thank you
Login to comment on this ticket.