When remote access fails, emit a better understandable error message, maybe "user XXX@demo.openwfm.org cannot access the directory YYY" instead of having to autopsy the traceback. Or, at least, as a part of the eror handling in ssh shuttle, print the relevant part of conf.json (remote user, remote directory)
When remote access fails, emit a better understandable error message, maybe "user XXX@demo.openwfm.org cannot access the directory YYY" instead of having to autopsy the traceback. Or, at least, as a part of the eror handling in ssh shuttle, print the relevant part of conf.json (remote user, remote directory)