OK, maybe I’ve been asking the wrong questions. I’ve been trying to figure out how to use Convoy/Gluster to take cluster-wide snapshotted backups using Gluster’s LVM support, but in doing research it seems that even snapshot backups require some level of technology-level integration. I.e. I still have to block databases from writing to their files while the snapshot is in progress.
So given you have data available across hosts, in something like Gluster, and maybe a variety of stacks running in your Rancher environment, what are your backup strategies? Is there a sidekick for every container with data you’d like to back up, that knows how to back up its specific data and runs on a regular schedule? Something else entirely? This problem is simple if you have data on a single host and can just write a cron job, but I haven’t found a serious treatment on how to back up data when you have a whole bunch of technologies, they’re scattered across hosts in a cluster, and all their data is scattered across the environment and isn’t necessarily on a single host for easy backup.
Thanks.