How do I get a 2nd drobo fs at a remote site to see the first drobo to back it up?
The main unit is on a dsl connection with a dyndns address and vpn. It has a fixed local address. I understand making rsync work on a local network, but how do I get the 2nd drobo to connect through the router.
You will need to port forward some ports through one or both routers.
If you are using rsync over ssh (which is what I would do) then you need to pick a port on the outside of your router and forward it to the SSH port on the Drobo FS you are trying to connect to. Try http://www.portforward.com to figure out how to do it.
The default SSH port is 22, but you don’t need to use that if you don’t want to. Some folks say that you should use another port so people find it hard to find your SSH connection to poke at it. You can forward any outside port to 22 on your FS.
So I’ll need Dropbear and rSync installed on both. Both will need their shares configured in the rSync config file (I’ve seen that a few places, I think I can do that correctly). Then I’ll need to do what to actually make them connect?
You can also pull from the destination instead of pushing from the source, it’s up to you.
Replace remote_user_name, remote_port, name.or_ip.of.destination and /name/of/*/directory as applicable.
If you add a -n while you are doing initial testing it will do a “dry run” and tell you what would be transferred but not actually transfer anything. This can be useful while you are messing witth getting the right path, etc.
You will need to enter the remote password. Once you can get this to work there are ways to set up public/private key based secure logins using SSH so you won’t need the passwords any more.
From what I remember the directory names on the FS are a bit odd (/mnt/shares/something like that). I’m sure there is info on here somewhere.
To make it easier you should do some sort of dynamic DNS so there is a DNS name for the destination. Try http://www.dyndns.com if you don’t already have a solution.
This was just a one-time test so that you could verify that everything could talk to each other. Once you get communication going then you would need to write some scripts and use cron or something like that to run them.
I do something a bit like this at work, so I can publish the scripts after I make them a bit more generic.