Rancher 2.x Backup - Maintenance Mode?

Rancher
1

We run a non-HA Rancher server, is there any kind of API to put Rancher 2.x into a “maintenance mode” in order to backup the data? The current backup strategy is just to backup the entire storage volume that is bind-mounted into the rancher-server container, however, if there much load at all, it seems like the etcd snap/db files fail to backup due to “file changed while we read it” errors.

If there were an API to temporarily put the Rancher server into some kind of “maintenance mode” during backups, that would be great (perhaps keep the k8s proxy capabilities up, but temporarily disable changes to Rancher?). We have requirements against stopping the server all together for a daily backup cron.

2

That’s not really very practical, there’s nodes reporting state all the time, UI live updating is driven by watches on the database, eventually hosts will be appearing missing to the controller because there’s no state updates and it will want to delete/replace them.

It would probably be better to follow the HA install but tweak it for only one node, then you get automatic etcd backups which can be copied out.

Or you could have the docker volume be on a filesystem that supports snapshotting and get an instant point-in-time copy of the single container’s disk.