Comment by thomasbacklund

Hey :)

Thanks for the tip! Yes, upping the site with articles and examples is on the TODO :)

In the SSH module doc [0] there are a few examples, such as:

  space -m ssh /ssh/ -e SSHHOST=jump1.example.com,jump2.example.com,jump3.example.com,destination.example.com -e SSHCOMMAND="ls -l"

One very fun feature is that you can tab auto-complete filenames and container names and also files inside container on a remote host:

  space -m ssh -e SSHHOST=1.2.3.4 -e SSHPORT=2233 -e SSHUSER=rabbit /wrap/ -m docker /wrap_exec/ -e DOCKERCONTAINER=[tab tab]

Then:

  space -m ssh -e SSHHOST=1.2.3.4 -e SSHPORT=2233 -e SSHUSER=rabbit /wrap/ -m docker /wrap_exec/ -e DOCKERCONTAINER=something -m file /cat/ -e file=/home/[tab tab]

Note that server jumping (over firewalls) is perfectly viable in the last example too, by adding more hosts (comma separated). Note that it can be important to balance host names to user names (see the doc [0]).

In the examples space wraps module functionality where the last module file is the one finally running inside the container, wrapped first by the `ssh` module then the `docker` module.

It is totally agent-less so nothing is uploaded to the server, it's just like SSHing in and running commands.

If you add the `-d` flag to space it will dump the script to output, which you can save to a .sh file and run later (without needing space.sh at all). Space it self requires bash, but the output runs with POSIX shells, so it also works with dash/ash/busybox, etc which is good because servers/containers often doesn't come with bash, but simpler ash/dash instead.

Thanks for showing interest, this makes me keen on pushing a few fixes and updates I have pending. Let me know if it doesn't work and I'll fix that :)

[0] https://space.sh/gitlab.com/space-sh/ssh/doc/master/examples...