Given that description, Cloudmin could help with the failover/recovery of my sites?
Yes, and no. Cloudmin does not yet address failover issues. It will…but it’s not the tool for the job today.
If I handle all the data sync, and tell Cloudmin about my two Virtualmin servers, Cloudmin will be able to move all my sites from Linode to FSCKVps if my Linode fails?
Somewhat, but not really.
Here’s what Cloudmin can do today:
Backup Virtualmin virtual servers automatically and copy the backup to another Virtualmin machine, which could then be restored on the other server, either via the Virtualmin UI or via a command run from the command line. This is actually possible without Cloudmin. Just setup Virtualmin on each server to make backups via ssh to the other server. All the work is happening in Virtualmin…Cloudmin just happens to make it slightly easier to navigate between the two machines.
Or, and this one requires Cloudmin:
When told to do so, Cloudmin can migrate a Virtualmin virtual server from one machine to another (via a backup/restore). This requires both machines to be running, however…it’s not intended (currently) to provide failover capabilities. It is entirely intended to make it easier to move Virtualmin virtual servers between machines in a Cloudmin cloud.
Both of these ignore the hard problems (the hard problems are databases, dynamic data that changes from minute to minute on your site, email spools, etc., all of which are not safe to just turn off and on on different machines at random because data will be lost during the time between backup and restore as well as the transition period to the failover and probably during the return to normal server).
All that said, it would obviously be a nice feature to have some sort of automatic recovery and failover. But, as I mentioned, the hard parts are not in our control. The easy stuff is easy, regardless of our support for it (which is why we’ve never spent a lot of time on it). The hard stuff is hard because it requires infrastructure that most of our customers simply don’t currently have (and the hard stuff simply cannot be solved satisfactorily between systems without high speed links between them).
We’re tinkering now with cluster filesystems, like Red Hat’s GFS, and I’ve also heard that replication has gotten much easier in recent versions of MySQL, so that’ll get us part of the way to a good solution that works on off-the-shelf hardware, and even (to some degree) across slow links between data centers like you’re working with.
So, basically, all I can really say is that we’re working on it.