I have a desktop running Linux Mint 18.3 and am learning how to use rsync for backing up. I have it working, mirroring data directories on my hard drive to another drive in a knockabout machine on my LAN. My next step is to get it to backup using SSH over the internet to a dedicated hard drive connected to a dedicated Raspberry Pi wired to a friend's router. I am trying to work out how to handle the IP address resolution as my friend (hosting the server) does not have a fixed IP, whereas I (hosting the client) DO have a fixed IP address from my ISP. It seems to me that a nifty way to avoid DNS issues is for the remote server to establish the SSH connection to my desktop which my client then uses for rsync. Is this feasible? Am I right to view the SSH connection as an armoured bi-directional pipe which can be used in this way, obviously dropping the ssh option from the rsync command? Are there issues or pitfalls I cannot spot in my inexperience?
Yes. You may think about this in overly complicated way. The backup server can use rsync to ssh to client and do the backup. rsync source destination the usual way. See the examples in rsync man page.
