#runlevels

How to use Systemd Timers on Linux

A systemd timer is a unit configuration file that provides a simple and effective way to schedule tasks on Linux. These timers offer an alternative to cron jobs and are the newer method of triggering jobs at specified times. Timer files, just like service files, are units controlled by systems. The following guide will demonstrate how to run a simple bash script, at regular intervals, using a…

runlevels are an operating mode a linux os runs in. there are usually 7 different runlevels

0 - halt, system shutdown 1 - single user mode, for administrative stuff, used if you want to reset the root password after you forgot it 2 - multi-user mode, like 'normal' mode, but without any network services running 3 - multi-user mode with networking, the normal operating mode of linux operating systems 4 - custom, a user defined mode that generally isn't used 5 - display manager, run level 3 plus a display manager 6 - reboot, does a restart of the operating system

how does the system know for each run level, what to do?

inside /etc/ you will see a bunch of directories named rc#.d like this

$ sudo ls -l /etc/rc* drwxr-xr-x. 2 root root 4096 Oct 13 03:41 init.d -rwxr-xr-x. 1 root root 2617 Aug 13 23:58 rc drwxr-xr-x. 2 root root 4096 Oct 13 03:41 rc0.d drwxr-xr-x. 2 root root 4096 Oct 13 03:41 rc1.d drwxr-xr-x. 2 root root 4096 Oct 13 03:44 rc2.d drwxr-xr-x. 2 root root 4096 Oct 13 03:44 rc3.d drwxr-xr-x. 2 root root 4096 Oct 13 03:44 rc4.d drwxr-xr-x. 2 root root 4096 Oct 13 03:44 rc5.d drwxr-xr-x. 2 root root 4096 Oct 13 03:41 rc6.d -rwxr-xr-x. 1 root root 220 Aug 13 23:58 rc.local -rwxr-xr-x. 1 root root 19841 Aug 13 23:58 rc.sysinit

these directories contain scripts the system uses at startup and shutdown to know in which order to start and stop applications and utilities for each of the 7 run levels.

run level 3 might have stuff like this

$ sudo ls -l /etc/rc3.d/ K01smartd -> ../init.d/smartd K09lwsmd -> ../init.d/lwsmd K10cups -> ../init.d/cups K10psacct -> ../init.d/psacct K10saslauthd -> ../init.d/saslauthd K15htcacheclean -> ../init.d/htcacheclean . . . K87rpcbind -> ../init.d/rpcbind K89rdisc -> ../init.d/rdisc K99rngd -> ../init.d/rngd S01sysstat -> ../init.d/sysstat S02lvm2-monitor -> ../init.d/lvm2-monitor S07iscsid -> ../init.d/iscsid S08ip6tables -> ../init.d/ip6tables S08iptables -> ../init.d/iptables S10network -> ../init.d/network . . . lrwxrwxrwx. 1 root root 15 Mar 24 2015 S97rhnsd -> ../init.d/rhnsd lrwxrwxrwx. 1 root root 19 Mar 24 2015 S97rhsmcertd -> ../init.d/rhsmcertd lrwxrwxrwx. 1 root root 19 Mar 30 2015 S98pe-puppet -> ../init.d/pe-puppet lrwxrwxrwx. 1 root root 11 Oct 9 17:04 S99local -> ../rc.local

these represent the kill order and start order for each run level. for startup, the scripts are executed in the order of S plus the lowest numbers in an alphabetical order, not numeric, so it would go something like

S01 -> S12 -> S1

and the kill order is the same

K01 -> K99

important systems services are started early on boot, and on shutdown are done towards the end, like a first in last out sequence.

when you use the chkconfig utility to add a service to start on boot or something, chkconfig looks inside /etc/init.d/yourservice in the comments at the top, for the line that specifies which runlevels and start order and kill order to apply, then it goes and sets up the links in these directories

#!/bin/sh # # supervisord Startup script for supervisord # # chkconfig: 2345 85 15 # processname: supervisord

in the snippet above on the chkconfig line, 2345 are the run levels, 85 is the start order, so quite late, and 15 is the kill order, so quite early

you can't create an init.d script without this information at the top, as it's important for the startup and shutdown sequences of the host

if you want to change runlevels, you can do something like this

$ sudo init 6

So, I'll begin a series of basic hacks; really simple foundations to start with.

First off: Single User Mode. 

Well, there are several ways to control a computer which you do not have access to. Assuming you have physical access of a machine, but do not have access to an account on the machine, single user (aka verbose) mode is probably the single most effective way to control it.

I won't go into much detail about what it is (read the wikipedia article for that), but it is simply a runlevel which allows one to control a computer as the superuser.

On Mac OS X it looks like this:

I

You can change the root password of the system and fiddle around with lots of important files like sudoers and pam.conf. 

How?

Reboot your computer. Hold down command-S (⌘-S). If your computer is not protected by an open firmware password, you should see something like the screen shown in the picture. It'll ask you to fsck with the command /sbin/fsck -fy and /sbin/mount -uw / to make modifications to the file system. 

Well, what do you want to do now that you have control over this computer?

Change the root password: Use the command passwd 

Add a user to the sudoers file: visudo -f /etc/sudoers (man visudo for more info)

Well, those are the two most useful commands, but go ahead and play around yourself to discover more. 

Limitations of Single User Mode:

Single user mode has several limitations. First of all, it is easily disabled. One can switch it off easily using the firmware password utility.