I have a Julia application running under Linux. I would like to make a couple of remote function calls to another Julia process running on a specific Windows machine, because that Windows machine has special data-acquisition hardware for which only a Win32 driver is available.
I have Julia installed on both the Linux and Windows machine, and I have Microsoft’s OpenSSH port set up such that I can perform password-free ssh logins from Linux to Windows. However, the Linux and Windows machines have of course very different file paths.
The Julia documentation on distributed computing focuses at the moment very much on clusters of worker machines that share the same file-path name space, and where the user may not care much on which exact worker machine a job ends up.
Where do I start if, rather than creating a homogeneous compute cluster, I just want to call Julia functions on one particular machine (to control some hardware there) where Julia and my code are installed on very different file paths (because its a rather different operating system)?
Is there somewhere a lower layer API than the ClusterManager that I can use to open just one ssh channel from a Linux machine to a Windows machine to call there a couple of Julia functions?