Add text talking about processes, daemons, and how to send them signals.

This commit is contained in:
Nik Clayton 2001-07-30 23:44:00 +00:00
parent 2fdac55c46
commit 7231511d66
Notes: svn2git 2020-12-08 03:00:23 +00:00
svn path=/head/; revision=10101

View file

@ -1,7 +1,7 @@
<!--
The FreeBSD Documentation Project
$FreeBSD: doc/en_US.ISO8859-1/books/handbook/basics/chapter.sgml,v 1.32 2001/07/25 20:41:24 chern Exp $
$FreeBSD: doc/en_US.ISO8859-1/books/handbook/basics/chapter.sgml,v 1.33 2001/07/30 17:18:49 murray Exp $
-->
<chapter id="basics">
@ -192,6 +192,295 @@
contains.</para>
</sect1>
<sect1 id="basics-processes">
<title>Processes</title>
<para>FreeBSD is a multi-tasking operating system. This means that it
seems as though more than one program is running at once. Each program
running at any one time is called a <firstterm>process</firstterm>.
Every command you run will start at least one new process, and there are
a number of system processes that run all the time, keeping the system
functional.</para>
<para>Each process is uniquely identified by a number called a
<firstterm>process ID</firstterm>, or <firstterm>PID</firstterm>, and,
like files, each process also has one owner and group. The owner and
group information is used to determine what files and devices the
process can open, using the file permissions discussed earlier. Most
processes also have a parent process. The parent process is the process
that started them. For example, if you are typing commands to the shell
then the shell is a process, and any commands you run are also
processes. Each process you run in this way will have your shell as its
parent process. The exception to this is a special process called
<command>init</command>. <command>init</command> is always the first
process, so its PID is always 1. <command>init</command> is started
automatically by the kernel when FreeBSD starts.</para>
<para>Two commands are particularly useful to see the processes on the
system, &man.ps.1; and &man.top.1;. The &man.ps.1; command is used to
show a static list of the currently running processes, and can show
their PID, how much memory they are using, the command line they were
started with, and so on. The &man.top.1; command displays all the
running processes, and updates the display every few seconds, so that
you can interactively see what your computer is doing.</para>
<para>By default, &man.ps.1; only shows you the commands that are running
and are owned by you. For example;</para>
<screen>&prompt.user; <userinput>ps</userinput>
PID TT STAT TIME COMMAND
298 p0 Ss 0:01.10 tcsh
7078 p0 S 2:40.88 xemacs mdoc.xsl (xemacs-21.1.14)
37393 p0 I 0:03.11 xemacs freebsd.dsl (xemacs-21.1.14)
48630 p0 S 2:50.89 /usr/local/lib/netscape-linux/navigator-linux-4.77.bi
48730 p0 IW 0:00.00 (dns helper) (navigator-linux-)
72210 p0 R+ 0:00.00 ps
390 p1 Is 0:01.14 tcsh
7059 p2 Is+ 1:36.18 /usr/local/bin/mutt -y
6688 p3 IWs 0:00.00 tcsh
10735 p4 IWs 0:00.00 tcsh
20256 p5 IWs 0:00.00 tcsh
262 v0 IWs 0:00.00 -tcsh (tcsh)
270 v0 IW+ 0:00.00 /bin/sh /usr/X11R6/bin/startx -- -bpp 16
280 v0 IW+ 0:00.00 xinit /home/nik/.xinitrc -- -bpp 16
284 v0 IW 0:00.00 /bin/sh /home/nik/.xinitrc
285 v0 S 0:38.45 /usr/X11R6/bin/sawfish</screen>
<para>As you can see in this example, the output from &man.ps.1; is
organized in to a number of columns. <literal>PID</literal> is the
process ID discussed earlier. PIDs are assigned starting from 1, go up
to 65536, and wrap around back to the beginning when you run out.
<literal>TT</literal> shows the tty the program is running on, and can
safely be ignore for the moment. <literal>STAT</literal> shows the
program's state, and again, can be safely ignored.
<literal>TIME</literal> is the amount of time the program has been
running on the CPU&mdash;this is not necessarily the elapsed time since
you started the program, as some programs spend a lot of time waiting
for things to happen before they need to spend time on the CPU.
Finally, <literal>COMMAND</literal> is the command line that was used to
run the program.</para>
<para>&man.ps.1; supports a number of different options to change the
information that is displayed. One of the most useful sets is
<literal>auxww</literal>. <option>a</option> displays information
about all the running processes, not just your own. <option>u</option>
displays the username of the process' owner, as well as memory usage.
<option>x</option> displays information about daemon processes, and
<option>ww</option> causes &man.ps.1; to display the full command line,
rather than truncating it once it gets too long to fit on the
screen.</para>
<para>The output from &man.top.1; is similar. A sample session looks like
this;</para>
<screen>&prompt.user; <userinput>top</userinput>
last pid: 72257; load averages: 0.13, 0.09, 0.03 up 0+13:38:33 22:39:10
47 processes: 1 running, 46 sleeping
CPU states: 12.6% user, 0.0% nice, 7.8% system, 0.0% interrupt, 79.7% idle
Mem: 36M Active, 5256K Inact, 13M Wired, 6312K Cache, 15M Buf, 408K Free
Swap: 256M Total, 38M Used, 217M Free, 15% Inuse
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
72257 nik 28 0 1960K 1044K RUN 0:00 14.86% 1.42% top
7078 nik 2 0 15280K 10960K select 2:54 0.88% 0.88% xemacs-21.1.14
281 nik 2 0 18636K 7112K select 5:36 0.73% 0.73% XF86_SVGA
296 nik 2 0 3240K 1644K select 0:12 0.05% 0.05% xterm
48630 nik 2 0 29816K 9148K select 3:18 0.00% 0.00% navigator-linu
175 root 2 0 924K 252K select 1:41 0.00% 0.00% syslogd
7059 nik 2 0 7260K 4644K poll 1:38 0.00% 0.00% mutt
...</screen>
<para>The output is split in to two sections. The header (the first five
lines) shows the PID of the last process to run, the system load averages
(which are a measure of how busy the system is), the system uptime (time
since the last reboot) and the current time. The other figures in the
header relate to how many processes are running (47 in this case), how
much memory and swap space has been taken up, and how much time the
system is spending in different CPU states.</para>
<para>Below that are a series of columns containing similar information
to the output from &man.ps.1;. As before you can see the PID, the
username, the amount of CPU time taken, and the command that was run.
&man.top.1; also defaults to showing you the amount of memory space
taken by the process. This is split in to two columns, one for total
size, and one for resident size&mdash;total size is how much memory the
application has needed, and the resident size is how much it is actually
using at the moment. In this example you can see that Netscape has
needed almost 30MB of RAM, and is currently only needing 9MB.</para>
<para>&man.top.1; automatically updates this display every two seconds;
this can be changed with the <option>s</option> option.</para>
</sect1>
<sect1>
<title>Daemons, signals, and killing processes</title>
<para>When you run an editor it is easy to control the editor, tell it to
load files, and so on. You can do this because the editor provides
facilities to do so, and because the editor is attached to a
<firstterm>terminal</firstterm>. Some programs are not designed to be
run with continuous user input, and so they disconnect from the terminal
at the first opportunity. For example, a web server spends all day
responding to web requests, it normally does not need any input from
you. Programs that transport e-mail from site to site are another
example of this class of application.</para>
<para>We call these programs <firstterm>daemons</firstterm>. Daemons were
characters in Greek mythology; neither good or evil, they were little
attendant spirits that, by and large, did useful things for mankind.
Much like the web servers and mail servers of today do useful things.
This is why the BSD mascot has, for a long time, been the cheerful
looking daemon with sneakers and a pitchfork.</para>
<para>There is a convention to name programs that normally run as daemons
with a trailing <quote>d</quote>. <application>BIND</application> is the
Berkeley Internet Name Daemon (and the program you can is called
<command>named</command>), the <application>Apache</application> web
server program is called <command>httpd</command>, the line printer
spooling daemon is <command>lpd</command> and so on. This is a
convention, not a hard and fast rule; for example, the main mail daemon
for the <application>Sendmail</application> application is called
<command>sendmail</command>, and not <command>maild</command>, as you
might imagine.</para>
<para>Sometimes you will need to communicate with a daemon process. These
commmunications are called <firstterm>signals</firstterm>, and you can
communicate with daemons (or with any running process) by sending it a
signal. There are a number if different signals that you can
send&mdash;some of them have a specific meaning, others are interpreted
by the application, and the application's documentation will tell you
how that application interprets signals. You can only send a signal to
a process that you own. If you try and send a signal to someone else's
process it will be ignored. The exception to this is the
<username>root</username> user, who can send signals to everyone's
processes.</para>
<para>FreeBSD will also send applications signals in some cases. If an
application is badly written, and tries to access memory that it is not
supposed to, FreeBSD sends the process the <firstterm>Segmentation
Violation</firstterm> signal (<literal>SIGSEGV</literal>). If an
application has used the &man.alarm.3; system call to be alerted after a
period of time has elapsed then it will be sent the Alarm signal
(<literal>SIGALRM</literal>), and so on.</para>
<para>Two signals can be used to stop a process,
<literal>SIGTERM</literal> and <literal>SIGKILL</literal>.
<literal>SIGTERM</literal> is the polite way to kill a process; the
process can <emphasis>catch</emphasis> the signal, realise that you want
it to shut down, close any log files it may have open, and generally
finish whatever it is doing at the time before shutting down. In some
cases a process may even ignore <literal>SIGTERM</literal> if it is in
the middle of some task that can not be interrupted.</para>
<para><literal>SIGKILL</literal> can not be ignored by a process. This is
the <quote>I don't care what you are doing, stop right now</quote>
signal. If you send <literal>SIGKILL</literal> to a process then
FreeBSD will stop that process there and then<footnote>
<para>Not quite true&mdash;there are a few things that can not be
interrupted. For example, if the process is trying to read from a
file that is on another computer on the network, and the other
computer has gone away for some reason (been turned off, or the
network has a fault), then the process is said to be
<quote>uninterruptible</quote>. Eventually the process will time
out, typically after two minutes. As soon as this time out occurs
the process will be killed.</para>
</footnote>.</para>
<para>The other signals you might want to use are
<literal>SIGHUP</literal>, <literal>SIGUSR1</literal>, and
<literal>SIGUSR2</literal>. These are general purpose signals, and
different applications will do different things when they are
sent.</para>
<para>Suppose that you have changed your web server's configuration
file&mdash;you would like to tell the web server to re-read its
configuration. You could stop and restart <command>httpd</command>, but
this would result in a brief outage period on your webserver, which may
be undesirable. Most daemons are written to respond to the
<literal>SIGHUP</literal> signal by re-reading their configuration
file. So instead of killing and restarting <command>httpd</command> you
would send it the <literal>SIGHUP</literal> signal. Because there is no
standard way to respond to these signals, different daemons will have
different behaviour, so be sure and read the documentation for the
daemon in question.</para>
<para>Signals are sent using the &man.kill.1; command, as this example
shows.</para>
<procedure>
<title>Sending a signal to a process</title>
<para>This example shows how to send a signal to &man.inetd.8;. The
&man.inetd.8; configuration file is
<filename>/etc/inetd.conf</filename>, and &man.inetd.8; will re-read
this configuration file when it is sent
<literal>SIGHUP</literal>.</para>
<step>
<para>Find the process ID of the process you want to send the signal
to. Do this using &man.ps.1; and &man.grep.1;. The &man.grep.1;
command is used to search through output, looking for the string you
specify. This command is run as a normal user, and &man.inetd.8; is
run as <username>root</username>, so the <option>ax</option> options
must be given to &man.ps.1;.</para>
<screen>&prompt.user; <userinput>ps -ax | grep inetd</userinput>
198 ?? IWs 0:00.00 inetd -wW</screen>
<para>So the &man.inetd.8; PID is 198. In some cases the
<literal>grep inetd</literal> command might also occur in this
output. This is because of the way &man.ps.1; has to find the list
of running processes.</para>
</step>
<step>
<para>Use &man.kill.1; to send the signal. Because &man.inetd.8; is
being run by <username>root</username> you must use &man.su.1; to
become <username>root</username> first.</para>
<screen>&prompt.user; <userinput>su</userinput>
<prompt>Password:</prompt>
&prompt.root; <userinput>/bin/kill -s HUP 198</userinput></screen>
<para>As is common with Unix commands, &man.kill.1; will not print any
output if it is successfully. If you try and send a signal to a
process that you do not own then you will see <errorname>kill:
<replaceable>PID</replaceable>: Operation not
permitted</errorname>. If you mistype the PID you will either
send the signal to the wrong process, which could be bad, or, if
you are lucky, you will have sent the signal to a PID that is not
currently in use, and you will see <errorname>kill:
<replaceable>PID</replaceable>: No such process</errorname>.</para>
<note>
<title>Why use <command>/bin/kill</command>?</title>
<para>Many shells provide the <command>kill</command> command as a
built in command; that is, the shell will send the signal
directly, rather than running <filename>/bin/kill</filename>.
This can be very useful, but different shells have a different
syntax for specifying the name of the signal to send. Rather than
try to learn all of them, it can be simpler just to use the
<command>/bin/kill <replaceable>...</replaceable></command>
command directly.</para>
</note>
</step>
</procedure>
<para>Sending other signals is very similar, just substitute
<literal>TERM</literal> or <literal>KILL</literal> in the command line
as necessary.</para>
<important>
<para>Killing random process on the system can be a bad idea. In
particular, &man.init.8;, process ID 1, is very special. Running
<command>/bin/kill -s KILL 1</command> is a quick way to shutdown your
system. <emphasis>Always</emphasis> double check the arguments you
run &man.kill.1; with <emphasis>before</emphasis> you press
<keycap>RETURN</keycap>.</para>
</important>
</sect1>
<sect1 id="shells">
<title>Shells</title>
<indexterm><primary>shells</primary></indexterm>