path: root/chapter_16.xml

<?xml version="1.0"?>
<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
  "/usr/share/xml/docbook/xml-dtd-4.5/docbookx.dtd">

<chapter>
<title>Basic Networking Utilities</title>

<para>
So you've finally managed to setup your network connection, now what?
How do you know that it's working? How do you know that you set it up
correctly? And just what do you do now that it's setup? Well this
chapter is for you.
</para>

<section>
<title>Network Diagnostic Tools</title>

<para>
Slackware Linux includes a great many networking tools for
troubleshooting and diagnosing network connection troubles, or just for
seeing what's out there on the network. Most of these tools are
command-line tools, so you can run them from a virtual terminal or in a
console window on your graphical desktop. A few of them even have
graphical front-ends, but we're going to deal almost exclusively with
command-line tools for now.
</para>

<section>
<title>ping</title>

<para>
<application>ping</application>(8) is a handy tool for determining if a
computer is operational on your network or on the Internet at large.
You can think of as a type of sonar for computers. By using it, you
send out a "ping" and listen for an echo to determine if another
computer or network device is listening. By default,
<application>ping</application> checks for the remote computer once per
second indefinitely, but you can change the interval between checks and
the total number of checks easily, just check the man page. You can
terminate the application at any time with
<keycap>CTRL</keycap>-<keycap>c</keycap>. When
<application>ping</application> is finished, it displays a handy
summary of its activity. <application>ping</application> is very useful
for determining if a computer on your network or the Internet is
available, but some systems block the packets
<application>ping</application> sends, so sometimes a system may be
functioning properly, but still not send replies.
</para>


<screen><prompt>darkstar:~# </prompt><userinput>ping -c 3 www.slackware.com</userinput>
64 bytes from slackware.com (64.57.102.34): icmp_seq=1 ttl=47 time=87.1 ms
64 bytes from slackware.com (64.57.102.34): icmp_seq=2 ttl=47 time=86.2 ms
64 bytes from slackware.com (64.57.102.34): icmp_seq=3 ttl=47 time=86.7 ms

--- slackware.com ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 86.282/86.718/87.127/0.345 ms
</screen>

</section>

<section>
<title>traceroute</title>

<para>
<application>traceroute</application>(8) is a handy tool for determining
what route your packets take to reach some other computer. It's mainly
of use for determining which computers are "near" or "far" from you.
This distance isn't strictly geographical, as your Internet Service
Provider may route traffic from your computer in strange ways.
<application>traceroute</application> shows you each router between
your computer and any other machine you wish to connect to.
Unfortunately, many providers, firewalls, and routers will block
<application>traceroute</application> so you might not get a complete
picture when using it. Still, it remains a handy tool for network
troubleshooting.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>traceroute www.slackware.com</userinput>
traceroute to slackware.com (64.57.102.34), 30 hops max, 46 byte
packets
 1  gw.ctsmacon.com (192.168.1.254)  1.468 ms  2.045 ms  1.387 ms
 2  10.0.0.1 (10.0.0.1)  7.642 ms  8.019 ms  6.006 ms
 3  68.1.8.49 (68.1.8.49)  10.446 ms  9.739 ms  7.003 ms
 4  68.1.8.69 (68.1.8.69)  11.564 ms  6.235 ms  7.971 ms
 5  dalsbbrj01-ae0.r2.dl.cox.net (68.1.0.142)  43.859 ms  43.287 ms
44.125 ms
 6  dpr1-ge-2-0-0.dallasequinix.savvis.net (204.70.204.146)  41.927 ms
58.247 ms  44.989 ms
 7  cr2-tengige0-7-5-0.dallas.savvis.net (204.70.196.29)  42.577 ms
46.110 ms  43.977 ms
 8  cr1-pos-0-3-3-0.losangeles.savvis.net (204.70.194.53)  78.070 ms
76.735 ms  76.145 ms
 9  bpr1-ge-3-0-0.LosAngeles.savvis.net (204.70.192.222)  77.533 ms
108.328 ms  120.096 ms
10  wiltel-communications-group-inc.LosAngeles.savvis.net
(208.173.55.186)  79.607 ms  76.847 ms  75.998 ms
11  tg9-4.cr01.lsancarc.integra.net (209.63.113.57)  84.789 ms  85.436
ms  85.575 ms
12  tg13-1.cr01.sntdcabl.integra.net (209.63.113.106)  87.608 ms
84.278 ms  86.922 ms
13  tg13-4.cr02.sntdcabl.integra.net (209.63.113.134)  87.284 ms
85.924 ms  86.102 ms
14  tg13-1.cr02.rcrdcauu.integra.net (209.63.114.169)  85.578 ms
85.285 ms  84.148 ms
15  209.63.99.166 (209.63.99.166)  84.515 ms  85.424 ms  85.956 ms
16  208.186.199.158 (208.186.199.158)  86.557 ms  85.822 ms  86.072 ms
17  sac-main.cwo.com (209.210.78.20)  88.105 ms  87.467 ms  87.526 ms
18  slackware.com (64.57.102.34)  85.682 ms  86.322 ms  85.594 ms
</screen>
</section>

<section>
<title>telnet</title>

<para>
Once upon a time, <application>telnet</application>(1) was the greatest
thing since sliced bread. Basically, <application>telnet</application>
opens an unencrypted network connection between two computers and hands
control of the session to the user rather than some other application.
Using <application>telnet</application>, people could connect to shells
on other computers and execute commands as if they were physically
present. Due to its unencrypted nature this is no longer recommended;
however, <application>telnet</application> is still used for this
purpose by many devices.
</para>

<para>
Today, <application>telnet</application> is put to better use as a
network diagnostic tool. Because it passes control of the session
directly to the user, it can be used for a great variety of testing
purposes. As long as you know what ASCII commands to send to the
receiving computer, you can do any number of activies, such as read web
pages or check your e-mail. Simply inform
<application>telnet</application> what network port to use, and you're
all set.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>telnet www.slackware.com 80</userinput>
Trying 64.57.102.34...
Connected to www.slackware.com.
Escape character is '^]'.
<userinput>HEAD / HTTP/1.1
Host: www.slackware.com
</userinput>
HTTP/1.1 200 OK
Date: Thu, 04 Feb 2010 18:01:35 GMT
Server: Apache/1.3.27 (Unix) PHP/4.3.1
Last-Modified: Fri, 28 Aug 2009 01:30:27 GMT
ETag: "61dc2-5374-4a973333"
Accept-Ranges: bytes
Content-Length: 21364
Content-Type: text/html
</screen>

</section>

<section>
<title>ssh</title>

<para>
As we mentioned, <application>telnet</application> may be useful as a
diagnostic tool, but its unencrypted nature makes it a security concern
for shell access. Thankfully, there's the secure shell protocol. Nearly
every Linux, UNIX, and BSD distribution today makes use of OpenSSH, or
<application>ssh</application>(1) for short. It is one of the most
commonly used network tools today and makes use of the strongest
cryptographic techniques. <application>ssh</application> has many
features, configuration options, and neat hacks, enough to fill its own
book, so we'll only go into the basics here.  Simply run
<application>ssh</application> with the user name and the host and
you'll be connected to it quickly and safely. If this is the first time
you are connecting to this computer, <application>ssh</application>
will ask you to confirm your desire, and make a local copy of the
encryption key to use. Should this key later change,
<application>ssh</application> will warn you and refuse to connect
because it is possible that some one is attempting to hijack the
connection using what is known as a man-in-the-middle attack.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>ssh alan@slackware.com</userinput>
alan@slackware.com's password: <userinput>secret</userinput>
<prompt>alan@slackware.com:~$ </prompt>
</screen>

<para>
The user and hostname are in the same form used by e-mail addresses.
If you leave off the username part, <application>ssh</application> will
use your current username when establishing the connection.
</para>

</section>

<section>
<title>tcpdump</title>

<para>
So far all the tools we've looked at have focused on making connections
to other computers, but now we're going to look at the traffic itself.
<application>tcpdump</application>(1) (which must be run as root)
allows us to view all or part of the network traffic originating or
received by our computer. <application>tcpdump</application> displays
the raw data packets in a variety of ways with all the network headers
intact. Don't be alarmed if you don't understand everything it
displays, <application>tcpdump</application> is a tool for professional
network engineers and system administrators. By default, it probes the
first network card it finds, but if you have multiple interfaces,
simply use the <arg>-i</arg> argument to specify which one you're
interested in.  You can also limit the data displayed using expressions
and change the manner in which it is displayed, but that is best
explained by the man page and other reference material.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>tcpdump -i wlan0</userinput>
tcpdump: verbose output suppressed, use -v or -vv for full protocol
decode
listening on wlan0, link-type EN10MB (Ethernet), capture size 96 bytes
13:22:28.221985 IP gw.ctsmacon.com.microsoft-ds > 192.168.1.198.59387:
Flags [P.], ack 838190560, win 3079, options [nop,nop,TS val 1382697489
ecr 339048583], length 164WARNING: Short packet. Try increasing the
snap length by 140
SMB PACKET: SMBtrans2 (REPLY)

13:22:28.222392 IP 192.168.1.198.59387 > gw.ctsmacon.com.microsoft-ds:
Flags [P.], ack 164, win 775, options [nop,nop,TS val 339048667 ecr
1382697489], length 134WARNING: Short packet. Try increasing the snap
length by 110
SMB PACKET: SMBtrans2 (REQUEST)
</screen>

</section>

<section>
<title>nmap</title>

<para>
Suppose you need to know what network services are running on a
machine, or multiple machines, or you wish to determine if multiple
machines are responsive?  You could <application>ping</application>
each one individually, <application>telnet</application> to each port
you're interested in, and note every detail, but that's very tedious
and time consuming. A much easier alternative is to use a port scanner,
and <application>nmap</application>(1) is just the tool for the job.
<application>nmap</application> is capable of scanning TCP and UDP
ports, determining the operating system of a network device, probing
each located service to determine its specific type, and much much
more.  Perhaps the simplist way to use <application>nmap</application>
is to "ping" multiple computers at once. You can use network address
notation (CIDR) or specify a range of addresses and
<application>nmap</application> will scan every one and return the
results to you when it's finished. You can even specify host names as
you like.
</para>

<para>
In order to "ping" hosts, you'll have to use the <arg>-sP</arg>
argument. The following command instructs
<application>nmap</application> to "ping" www.slackware.com and the 16
IP addresses starting at 72.168.24.0 and ending at 72.168.24.15.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>nmap -sP www.slackware.com 72.168.24.0/28</userinput>
</screen>

<para>
Should you need to perform a port scan, <application>nmap</application>
has many options for doing just that. When run without any arguments,
<application>nmap</application> performs a standard TCP port scan on all
hosts specified. There are also options to make
<application>nmap</application> more or less aggressive with its
scanning to return results quicker or fool intrusion detection
services. For a full discussion, you should refer to the rather
exhaustive man page. The following three commands perform a regular
port scan, a SYN scan, and a "Christmas tree" scan.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>nmap www.example.com</userinput>
<prompt>darkstar:~# </prompt><userinput>nmap -sS www.example.com</userinput>
<prompt>darkstar:~# </prompt><userinput>nmap -sX www.example.com</userinput>
</screen>

<para>
Be warned! Some Internet Service Providers frown heavily on port
scanning and may take measures to prevent you from doing it.
<application>nmap</application> and applications like it are best used
on your own systems for maintenance and security purposes, not as
general purpose Internet scanners.
</para>

</section>

</section>

<section><title>Web Browsers</title>

<para>
Slackware includes a variety of web browsers. If you're using a
graphical desktop, you'll find Firefox, Seamonkey, and others you may
already be familiar with, but what about console access? Fortunately,
there are a number of capable web browsers here as well.
</para>

<section><title>lynx</title>

<para>
The oldest console-based web browser included with Slackware is
definitely <application>lynx</application>(1), a very capable if
somewhat limited web browser.  <application>lynx</application> does not
support frames, javascript, or pictures; it is strictly a text web
browser. Navigation is performed using your keyboard's arrow keys and
optionally, a mouse. While it lacks many features that other browsers
support, <application>lynx</application> is one of the fastest web
browsers you'll ever use for gathering information. For example, the
<arg>-dump</arg> argument sends the formatted web page directly to the
console, which can then be piped to other programs.
</para>

<imagedata fileref="img/lynx.png" format="PNG"/>

</section>

<section><title>links</title>

<para>
A more feature-rich alternative is the popular
<application>links</application>(1), a console-based web browser that
supports frames and has better table rendering than
<application>lynx</application>. Like its predecessor,
<application>links</application> is navigated with the arrow keys, and
the use of a mouse is supported.  Unlike,
<application>lynx</application> it also includes a handy menu (simply
click on the top line with your mouse to activate) and generally
formats web pages better.
</para>

<imagedata fileref="img/links.png" format="PNG"/>

</section>

<section><title>wget</title>

<para>
Unlike the other browsers we've looked at,
<application>wget</application>(1) is non-interactive. Rather than display
HTTP content, <application>wget</application> downloads it. This takes
the "browsing" out of the web browser. Unlike the dump modes of other
browsers, <application>wget</application> does not format its
downloads; rather it copies the content in its exact form on the web
server with all tags and binary data in place. It also supports several
recursive options that can effectively mirror online content to your
local computer.  <application>wget</application> need not operate
exclusively on HTTP content; it also supports FTP and several other
protocols.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>wget ftp://ftp.osuosl.org/pub/slackware/slackware-current/ChangeLog.txt</userinput>
--2010-05-01 13:51:19--
ftp://ftp.osuosl.org/pub/slackware/slackware-current/ChangeLog.txt
           => `ChangeLog.txt'
Resolving ftp.osuosl.org... 64.50.236.52
Connecting to ftp.osuosl.org|64.50.236.52|:21... connected.
Logging in as anonymous ... Logged in!
==> SYST ... done.    ==> PWD ... done.
==> TYPE I ... done.  ==> CWD /pub/slackware/slackware-current ...  done.
==> SIZE ChangeLog.txt ... 75306
==> PASV ... done.    ==> RETR ChangeLog.txt ... done.
Length: 75306 (74K)

100%[======================================>] 75,306       110K/s   in 0.7s    

2010-05-01 13:51:22 (110 KB/s) - `ChangeLog.txt' saved [75306]
</screen>

</section>

</section>

<section>
<title>FTP Clients</title>

<para>
Lots of data is stored on FTP servers the world over. In fact,
Slackware Linux was first publically offered via FTP and continues to
be distributed in this fashion today. Most open source software can be
downloaded in source code or binary form via FTP, so knowing how to
retrieve this information is a handy skill.
</para>

<section><title>ftp</title>

<para>
The simplest FTP client included with Slackware is named simply,
<application>ftp</application>(1) and is a reliable if somewhat simple
means of sending and retrieving data. <application>ftp</application>
connects to an FTP server, asks for your username and password, and
then allows you to put or get data to and from that server.
<application>ftp</application> has fallen out of favor with more
experienced users do to a lack of features, but remains a handy tool,
and much of the documentation you see online will refer you to it.
</para>

<para>
Once an FTP session has been initialized, you'll be placed at a prompt
somewhat like a shell.  From here you can change and list directories
using the "cd" and "ls" commands, just like a shell.  Additionally, you
may issue the "put" command to send a file to the server, or a "get"
command to retrieve data from the server. If you're connecting to a
public FTP server, you'll want to use the "anonymous" username and
simply enter your e-mail address (or a fake one) for the password.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>ftp ftp.osuosl.org</userinput>
Name (ftp.osuosl.org:alan): <userinput>anonymous</userinput>
331 Please specify the password.
Password: <userinput>secret</userinput>
230 Login successful.
Remote system type is UNIX.
Using binary mode to transfer files.
ftp> <userinput>cd pub/slackware/slackware-current/</userinput>
250 Directory successfully changed.
ftp> <userinput>get ChangeLog.txt</userinput>
local: ChangeLog.txt remote: ChangeLog.txt
200 PORT command successful. Consider using PASV.
150 Opening BINARY mode data connection for ChangeLog.txt (33967
bytes).
226 File send OK.
33967 bytes received in 0.351 secs (94 Kbytes/sec)
ftp> <userinput>bye</userinput>
221 Goodbye.
</screen>

</section>

<section><title>ncftp</title>

<para>
<application>ncftp</application>(1) (pronounced nick-f-t-p), is a more
feature rich successor to <application>ftp</application>, supporting
tab completion and recursive retrieval. It automatically connects to a
server as the anonymous user, unless you specify a different username
on the commandline with the <arg>-u</arg> argument. The primary
advantage over <application>ftp</application> is the ability to send
and retrieve multiple files at once with the "mput" and "mget"
commands. If you pass the <arg>-R</arg> argument to either of them,
they will recursively put or get data from directories.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>ncftp ftp.osuosl.org</userinput>
Logging in...                                                                   
Login successful.
Logged in to ftp.osuosl.org.                                                    
ncftp / > <userinput>cd pub/slackware/slackware-current</userinput>
Directory successfully changed.
ncftp ...ware/slackware-current > <userinput>mget -R isolinux</userinput>
isolinux/README.TXT:                                     4.63 kB 16.77 kB/s  
isolinux/README_SPLIT.TXT:                             788.00 B    5.43 kB/s  
isolinux/f2.txt:                                       793.00 B    5.68 kB/s  
isolinux/initrd.img:                                    13.75 MB 837.91 kB/s  
isolinux/iso.sort:                                      50.00 B  354.50 B/s   
isolinux/isolinux.bin:                                  14.00 kB 33.99 kB/s  
isolinux/isolinux.cfg:                                 487.00 B    3.30 kB/s  
isolinux/message.txt:                                  760.00 B    5.32 kB/s  
isolinux/setpkg:                                         2.76 kB 19.11 kB/s  
ncftp ...ware/slackware-current > <userinput>bye</userinput>
</screen>

</section>

<section><title>lftp</title>

<para>
The last client we're going to look at is
<application>lftp</application>(1).  Like
<application>ncftp</application>, it supports tab completion and
recursive activity, but has a more friendly license.  Rather than user
"mget" and "mput", all recursive operations are handled with the
"mirror" command.  "mirror" has many different options available, so
I'll have to refer you to the man page and the built-in "help" command
for complete details.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>lftp ftp.osuosl.org</userinput>
lftp ftp.osuosl.org:~> <userinput>cd /pub/slackware/slackware-current</userinput>
cd ok, cwd=/pub/slackware/slackware-current            
lftp ftp.osuosl.org:/pub/slackware/slackware-current> <userinput>mirror isolinux</userinput>
Total: 2 directories, 16 files, 1 symlink                                      
New: 16 files, 1 symlink
14636789 bytes transferred in 20 seconds (703.7K/s)
lftp ftp.osuosl.org:/pub/slackware/slackware-current> <userinput>bye</userinput>
</screen>

</section>

</section>

<section>
<title>NNTP Clients</title>

<para>
Once upon a time when the Internet was young, before the World Wide Web
was invented and no one had heard of hyperlinks, everyone retrieved
their news and information through a service known as Usenet using the
NNTP protocol. It remains today a useful knowledge base of information
on an incredible variety of subjects, but if you wish to access this
information, you're going to need a proper client.  Slackware includes
a number of NNTP clients with both console and graphical interfaces,
but we'll only detail the console tools here. Popular graphical news
readers include <application>knode</application> and
<application>pan</application>.
</para>

<section><title>tin</title>

</section>

<section><title>slrn</title>

</section>

</section>

<section>
<title>rsync</title>

<para>
Ready to see something cool? Have you ever found yourself needing just
a handful of files from a large directory, but you're not entirely sure
which files you already have and which ones you need? You can download
the entire directory again, but that's duplicating a lot of work. You
can pick and chose, manually check everything, but that's very tedious.
Perhaps you've downloaded a large file such as an ISO, but something
went wrong with the download? It doesn't make sense that you should
have to pull down the entire file again if only a few bits have been
corrupted.  Enter <application>rsync</application>(1), a fast and
versatile copying tool for local and remote files.
</para>

<para>
<application>rsync</application> uses a handful of simple, but very
effective techniques to determine what needs to be changed. By checking
file size and time stamps, it can determine if two files are different.
If something has changed, it can determine what bytes are different,
and simply download that handfull of data rather than an entire file.
It is truly a marvel of modern technology.
</para>

<para>
In its simplist form, <application>rsync</application> connects to an
rsync protocol server and downloads a list of files and directories,
along with their sizes, timestamps, and other information. It then
compares this to the local files (if any) to determine what it needs to
transfer. Only files that are different will be synced.  Additionally,
it breaks up large files into smaller chunks and compares those chunks
using a quick and simple hash function. Any chunks that match are not
transferred, so the amount of data that must be copied can be
dramatically reduced.  <application>rsync</application> also supports
compression, verbose output, file deletion, permission handling, and
many other options. For a complete list, you'll need to refer to the
man page, but I've included a small table of some of the more common
options.
</para>

<table pgwide="0">
<title>rsync Arguments</title>
<tgroup cols="2">
  <thead>
    <entry>Argument</entry>
    <entry>Explaination</entry>
  </thead>
  <tbody>
    <row>
      <entry>-v</entry>
      <entry>Increased verbosity</entry>
    </row>
    <row>
      <entry>-c</entry>
      <entry>Checksum all files rather than relying on file size and timestamp</entry>
    </row>
    <row>
      <entry>-a</entry>
      <entry>Archive mode (equivilant to -rlptgoD)</entry>
    </row>
    <row>
      <entry>-e</entry>
      <entry>Specify a remote shell to use</entry>
    </row>
    <row>
      <entry>-r</entry>
      <entry>Recursive mode</entry>
    </row>
    <row>
      <entry>-u</entry>
      <entry>Update - skip files that are newer on the receiving end</entry>
    </row>
    <row>
      <entry>-p</entry>
      <entry>Preserve permissions</entry>
    </row>
    <row>
      <entry>-n</entry>
      <entry>Dry-run - perform a trial run without making any changes</entry>
    </row>
    <row>
      <entry>-z</entry>
      <entry>Compress - handy for slow network connections</entry>
    </row>
  </tbody>
</tgroup>
</table>

<para>
Due to the power and versatility of <application>rsync</application>,
it can be invoked in a number of ways.  The following two examples
connect to an rsync protocol server to retrieve some information and to
another server via ssh to encrypt the transmission.
</para>

<screen><prompt>darkstar:~# </prompt><userinput>rsync -avz rsync://ftp.osuosl.org/pub/slackware/slackware-current/ \
/src/slackware-current/</userinput>
<prompt>darkstar:~# </prompt><userinput>rsync -e ssh ftp.slackware.com:/home/alan/foo /tmp/foo</userinput>
</screen>

</section>

</chapter>