Different uses of SSH keys should be clarified

[markstos]

There are two places in the interface where I can put in SSH keys.

One is under /Configuration Files/.

The other is a "Private SSH" key for a specific server. It's particularly clear how these two interact, which ones becomes the default key, and when to put an SSH key in one place vs the other.

[markstos]

I found the SSH behavior could be debugged by temporarily replacing my deployment script with a "hello world" SSH command, with a verbose debugging turned on:

 ssh -vvv -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null user@host.com 'echo "Hello World"'

The options address the challenge of getting the "unknown host" prompt in the non-interactive context, and the -vvv flags supply verbose debugging output.

They show that first the the SSH key from "Configuration Files" is offered, then then "private key" set up for this particular server deployment is offered if the first key failed.

Once the basic SSH connect worked, then I could incorporate SSH into a more complex command.

[markoa]

Thanks for the feedback Mark! Can you explain what is it that you wanted to accomplish? I think it'd help understand one possible angle of (re)writing the related pages.

[markstos]

The final step of a deployment was to rsync-over-ssh, and I needed to get the related public-key authentication to work over SSH.