Re: screen refusing to reattach

[Jeremy Chadwick]

I can confirm this behaviour.  I've experienced it in BSD, Solaris, and
Linux.  I believe it has something to do with the tty already being
allocated somehow.

The solution is to kill off the shell process which had previously
run (and in some cases, still is running) screen -x or screen -r.  I
wrote a simple sh script called "cleartty" which does this for me
(since I myself only run one screen session at a time).

I've also seen behaviour when reattaching (screen -x or -r) works, then
a short bit later, stops working altogether.  I believe this to be
OpenSSH saying "Okay, the attached SSH session has died, lets clean up
applicable ttys/ptys" and proceeds to munge something that a reattached
screen seems to rely on.  Detatching, logging out, and logging back in
usually fixes it.

Authors' comments?

-- 
| Jeremy Chadwick                                 jdc at parodius.com |
| Parodius Networking                        http://www.parodius.com/ |
| UNIX Systems Administrator                   Mountain View, CA, USA |
| Making life hard for others since 1977.                             |

On Fri, Mar 25, 2005 at 09:45:18AM -0500, Phil!Gregory wrote:
> * Eric Smith <address@hidden> [2005-03-25 15:23 +0100]:
> > Ok, this has caused me enough inconvenience so I will finally ask about
> > it.  Every now and then screen freezes and if I try to reattach the
> > seassion elsewhere, the command `screen -x' just hangs.
> > 
> > Sometimes killing off the child processes seems to work, other times
> > not.  Usually after many hours the screen of its own accord becomes
> > re-attachable.
> 
> This sounds similar to a problem that I've run into.
> 
> When I've had a problem with screen along these lines, it's been waiting
> for blocked I/O, usually from an ssh session dying, but the system not
> having noticed (so it keeps sending data, but never gets a response).  As
> various send buffers start filling up, first the window that was active in
> that ssh session and then the entire screen session goes unresponsive.  If
> I wait long enough, the ssh session times out, and all of the blocking I/O
> returns, restoring the screen session to usability.
> 
> I'm not sure what advice to give, though.  I've tried setting 'nonblock
> on' in my .screenrc, as the man page seemed to imply that it might help,
> but it hasn't.  It's possible that it might work better for you.
> 
> -- 
> ...computer contrarian of the first order... / http://aperiodic.net/phil/
> PGP: 026A27F2  print: D200 5BDB FC4B B24A 9248  9F7A 4322 2D22 026A 27F2
> --- --
> Satellite Safety Tip #14:
>   If you see a bright streak in the sky coming at you, duck.
> ---- --- --
> 
> 
> _______________________________________________
> screen-users mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/screen-users