I want to post a summary on this, in case somebody finds it useful. My setup is the following: I have a notebook that connects to a server via ssh. The notebook is put to hibernate (suspended to disk) every so often, for at least 15 minutes at a time. The server is running nonstop. When put to hibernate, the notebook could possibly have several ssh connections open with the server, every one of them running separate applications, like editors, debuggers, etc.
The problem: How to arrange things so that when the notebook is brought back to life, the ssh connections are resumed, and the applications that they were running, resumed at the same point where they were when the hibernation started. The solution: Use a combination of autossh and screen. I modified the rscreen script shipped with autossh, so that the line that invokes autossh reads as follows: autossh -M -t $1 "screen -D -R -S Session-$NEXT" The reason why I added the -S option is because, without it, on thawing the notebook autossh would create new screen sessions on the server, leaving the interrupted ones still running, but detached. The NEXT variable is increased by 1 everytime rscreen is onvoked. I am not claiming that this is perfect, or even the right way to do it. It just seems to work for me though. Thanks so much to everybody who contributed to this thread. _______________________________________________ screen-users mailing list screen-users@gnu.org http://lists.gnu.org/mailman/listinfo/screen-users
