Description of problem: 'man telinit' says: .... or 1 to bring the system down into single-user mode AFAICS after 'telinit 1' most of things which were running are still running. A rough check with scripts in /etc/init.d/ quickly gives: acpid (pid 1958) is running... anacron (pid 2223) is running... atd (pid 2048) is running... auditd (pid 1861) is running... automount (pid 1943) is running... Avahi daemon is running crond (pid 2030) is running... cupsd (pid 2071) is running... gpm (pid 2023) is running... hald (pid 2078) is running... dbus-daemon (pid 1911) is running... rpc.statd (pid 1802) is running... ntpd (pid 1981) is running... privoxy (pid 2016) is running... rpcbind (pid 1785) is running... rpc.idmapd (pid 1836) is running... rsyslogd (pid 1925) is running... sendmail (pid 2000) is running... sm-client (pid 2008) is running... smartd (pid 2245) is running... sshd (pid 4048 4046 3287 1967) is running... xfs (pid 2042) is running... xinetd (pid 1974) is running... It is true that I cannot login as a non-root on a local console; but there is nothing which stops me doing that over a network. That is a mighty strange "single-user mode". One also can read in 'man telinit' about "S or s" that "you probably won’t want that". The only difference I can see between 'telinit 1' and 'telinit s' that in the first case "Tellinig INIT to go to single user mode." is printed, but not when the second form is used Version-Release number of selected component (if applicable): upstart-0.3.9-9.fc9 How reproducible: always
This sounds like a difference between runlevel configuration in Ubuntu and Fedora? In Ubuntu (and Debian) rc1.d contains K scripts for each of the services found in rc2-5, and the final script which issues "telinit S" to actually put the system into single-user mode
Fedora's the same way.
Will be fixed in initscripts-8.69-1. http://git.fedorahosted.org/git/?p=initscripts.git;a=commit;h=d61552366c960cae6f5a52d493185ec8d8e716cd
*** Bug 441489 has been marked as a duplicate of this bug. ***
> Will be fixed in initscripts-8.69-1. Well, with initscripts-8.69-1 installed I still see after 'telinit 1' '/usr/sbin/console-kit-daemon' and 'anacron -s' running. The second, in particular, is somewhat disconcerting as it is not obvious what it may do and when. Also what on earth is the following process hanging there: root 3072 0.0 0.2 11072 1096 tty1 Ss 12:59 0:00 /bin/sh -e -c ?runlevel --set S >/dev/null || true??/bin/bash???runlevel=$(/bin/awk -F ':' '$3 == "initdefault" { print $2 }' /etc/inittab)??[ -z "$runlevel" ] && runlevel="3"??telinit $runlevel? /bin/sh S All of this is surely in a much better shape (thanks!) but is it truly fixed?
(In reply to comment #5) > > Will be fixed in initscripts-8.69-1. > > Well, with initscripts-8.69-1 installed I still see after 'telinit 1' > '/usr/sbin/console-kit-daemon' and 'anacron -s' running. The second, > in particular, is somewhat disconcerting as it is not obvious what > it may do and when. Those would be bugs in the particular packages. > Also what on earth is the following process hanging there: > > root 3072 0.0 0.2 11072 1096 tty1 Ss 12:59 0:00 /bin/sh -e -c > ?runlevel --set S >/dev/null || true??/bin/bash???runlevel=$(/bin/awk -F ':' '$3 > == "initdefault" { print $2 }' /etc/inittab)??[ -z "$runlevel" ] && > runlevel="3"??telinit $runlevel? /bin/sh S That's the thing that starts the shell. Sort of a consequence as to how upstart executes it.
> Those would be bugs in the particular packages. In that case I better make sure that corresponding reports are filed.
Note that this only really applies to things started by init scripts; it's impossible to track everything that may have been started at some point.