Soren, I just now re-read your original description, and I see this:
debug1: Killed by signal 15.
That suggests that your *client* ssh is being killed.
The sshd debug supports that:
debug3: fd 9 is O_NONBLOCK
debug1: Setting controlling tty using TIOCSCTTY.
Received disconnect from 10.115.0.105: 11: disconnected by user
Is it possible there is something on your local machine killing ssh's for some reason? You said they get killed at the same exact time.. maybe a cron job?
Also re the pid file being present but ssh in stop/waiting.. if that pid is running sshd, then you have hit bug #531912 by accidentally using /etc/init.d/ssh instead of the upstart start/stop commands.
Soren, I just now re-read your original description, and I see this:
debug1: Killed by signal 15.
That suggests that your *client* ssh is being killed.
The sshd debug supports that:
debug3: fd 9 is O_NONBLOCK
debug1: Setting controlling tty using TIOCSCTTY.
Received disconnect from 10.115.0.105: 11: disconnected by user
Is it possible there is something on your local machine killing ssh's for some reason? You said they get killed at the same exact time.. maybe a cron job?
Also re the pid file being present but ssh in stop/waiting.. if that pid is running sshd, then you have hit bug #531912 by accidentally using /etc/init.d/ssh instead of the upstart start/stop commands.