> This is the typical way to use screen: > > - Turn on your laptop, open ssh connection to the server > - On the server: type 'screen' > - screen will give you a shell, just like when you log in normally > - Do work that you need inside that shell > - If you get disconnected, such as when you turn off your laptop, screen > will still be running on the server > - When you want to reconnect, turn on your laptop, and connect through > ssh to the server again > - type 'screen -D -R' to resume screen where it was before you turned > off your laptop. > > It is not possible to resume the exact SSH connection when you turn off > your laptop because networking doesn't work that way. Instead, you must > create a new SSH connection and then use screen to resume what you were > doing inside the screen session.
Thanks so much for the very useful feedback that you and other contributors to this list have come up with. It seems to be the case that using screen the way you describe, plus autossh, as pointed out by somebody else, goes a long way to achieving what I want. There is a wrinkle though: In my notebook, I usually have more than one ssh sessions to the server, on separate windows. If I use twice the rscreen script that comes with autossh, I can get two windows but attached to the same screen. That is, whatever I do in one of them, is just echoed on the other. This can be potentially useful, but it is not what I need. I would need to be able to open several, independent screen sessions on different windows, and automatically reattach to each of them correctly and separately after bringing the notebook back online again. Is this possible with a combination of autossh and screen? _______________________________________________ screen-users mailing list [email protected] http://lists.gnu.org/mailman/listinfo/screen-users
