[SOLVED] How to work on Debian Squeeze in runlevel 1 over ssh?

catkin · 05-27-2012, 12:20 AM

In runlevel 1, sshd is not running by default so something has to be done to work remotely via ssh. The advice used to be to remove the K*ssh links from /etc/rc{2..5}.d but there are no such links on Squeeze:

Code:

root@LS1:~# ll /etc/rc*.d/*ssh*
lrwxrwxrwx 1 root root 13 Mar 16  2011 /etc/rc2.d/S03ssh -> ../init.d/ssh
lrwxrwxrwx 1 root root 13 Mar 16  2011 /etc/rc3.d/S03ssh -> ../init.d/ssh
lrwxrwxrwx 1 root root 13 Mar 16  2011 /etc/rc4.d/S03ssh -> ../init.d/ssh
lrwxrwxrwx 1 root root 13 Mar 16  2011 /etc/rc5.d/S03ssh -> ../init.d/ssh

I tried an alternative solution, adding an ssh start link to /etc/rc1.d:

Code:

root@LS1:~# cd /etc/rc1.d && ln -s ../init.d/ssh S07ssh

but it did not work (trying to ssh in got "connection refused".

Unexpectedly, manually running /etc/init.d/ssh start on the console did enable ssh login. I do not understand why the init script link was not equivalent to the command.

Incidentally, using telinit 2 to leave runlevel 1 resulted in a hung system displaying "Starting NFS kernel daemon: nfsd". This was solved by using the power button to initiate shutdown and boot -- not an option when working remotely. The workaround is to use telinit 6 (not yet tested).

druuna · 05-27-2012, 01:56 AM

Hi,

If I add an ssh link to /etc/rc1.d it works, but only if the link has the correct S number. ../init.d/ssh needs to be started before ../init.d/single is started (have a look at /etc/init.d/single).

Your setup looks a bit different then mine (S03ssh vs S20ssh) , but I guess you get the idea:

Code:

# initial setup, no ssh:
# ls -l /etc/rc1.d/S*
lrwxrwxrwx 1 root root 19 May 20 14:20 /etc/rc1.d/S01killprocs -> ../init.d/killprocs
lrwxrwxrwx 1 root root 18 May 20 14:50 /etc/rc1.d/S23bootlogs -> ../init.d/bootlogs
lrwxrwxrwx 1 root root 16 May 20 14:50 /etc/rc1.d/S24single -> ../init.d/single

# new setup with ssh added:
# ls -l /etc/rc1.d/S*
lrwxrwxrwx 1 root root 19 May 20 14:20 /etc/rc1.d/S01killprocs -> ../init.d/killprocs
lrwxrwxrwx 1 root root 18 May 20 14:50 /etc/rc1.d/S23bootlogs -> ../init.d/bootlogs
lrwxrwxrwx 1 root root 15 May 27 08:46 /etc/rc1.d/S23ssh -> /etc/init.d/ssh
lrwxrwxrwx 1 root root 16 May 20 14:50 /etc/rc1.d/S24single -> ../init.d/single

After a reboot into RL 1, ssh will be up and running (it does on my box). If the ssh link is below the single link ssh will not be started due to what the single script starts (init -t1 S).

Quote:

Incidentally, using telinit 2 to leave runlevel 1 resulted in a hung system displaying "Starting NFS kernel daemon: nfsd"

This is mentioned in the init man page:

Code:

WARNINGS
On a Debian system, entering runlevel 1 causes all processes to be killed except for kernel threads and the
script  that  does the killing and other processes in its session.  As a consequence of this, it isn't safe
to return from runlevel 1 to a multi-user runlevel: daemons that were started in runlevel S and are  needed
for normal operation are no longer running.  The system should be rebooted.

catkin · 05-27-2012, 02:32 AM

Thanks drunna

That all makes perfect sense.