doc/en_US.ISO8859-1/articles/console-server/article.sgml
Tom Rhodes d4b003e30a FreeBSD documentation project prefers to use manual pages over man pages,
hence some documents should reflect this.

Tossed around on: -doc -developers
2002-07-11 19:08:31 +00:00

1231 lines
51 KiB
Text

<!-- $FreeBSD$ -->
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<title>Console Server</title>
<link rel="stylesheet" href="./itgastyle.css">
</head>
<body>
<h1>Console Server</h1>
<h2>The Problem</h2>
<p>You have a computer room with lots of Unix server machines and
lots of comms hardware. Each of these machines needs a serial
console. But serial terminals are hard to find and quite
expensive (especially compared to a much more capable PC). And
they take up a lot of precious space in the computer room.</p>
<p>You need access to the console because when things break, that
is where error messages go. And some tasks have to be done on
the console (e.g. boot problems or OS installs/upgrades). Some
Unix systems allow the console to break out to the ROM monitor
which can sometimes be the only way to unstick a hung machine.
This is often done with a LINE BREAK sent on the console serial
port.</p>
<p>If we are going to play about with consoles, then there are a
couple of other things that would be great:</p>
<ul>
<li>
<p>Remote access. Even in the same office, it would be
convenient to access all the consoles from your desk without
walking into the computer room. But often the machines are
off-site, perhaps even in another country.</p></li>
<li>
<p>Logging. If something has gone wrong, you would like to be
able to have a look at the previous console output to see
what is up. Ordinary console screens give you the last 25
lines. More would be better.</p>
<li>
<p>Network Independence. The solution needs to work even if the
network is down. After all, a failed network is when you need
consoles the most! Even better is network independence with
remote access.</p></li>
<li>
<p>No single-point failure. A console system that crashes every
machine when it fails is no use. This is particularly tricky
with Sun Unix hosts as they will interpret a powered-off
terminal as a BREAK, and drop back to the ROM monitor.</p></li>
<li>
<p>Interface with a pager or some similar alerter
device.</p></li>
<li>
<p>Ability to power-cycle machines remotely.</p></li>
<li>
<p>Not be <em>too</em> expensive. Free is even better!</p></li>
</ul>
<h2>Possible Solutions</h2>
<p>If you use PC hardware for your servers, then a so-called
"<em>KVM switch</em>" is one possible solution. This allows the
use of a single <em>K</em>eyboard, <em>V</em>ideo screen and
<em>M</em>ouse for multiple boxes. This cuts down on the space
problem, but only works for PC hardware (not any comms gear you
might have), and is not accessible from outside the computer
room. Nor does it have much scroll-back or logging, and you
have to handle alerting some other way. The big downside is
that it will not work for serial-only devices, such as
communications hardware. This means that even with a room full
of PC-based servers, you are probably still going to need some
sort of serial console solution. [Actually, Doug Schache has
pointed out that you <em>can</em> get KVM switches that also do
serial consoles or Sun-compatible KVM switching as well as PCs,
but they are expensive. See <a
href="http://www.avocent.com/">Avocent</a> for example.]</p>
<p>You might be tempted to do without a console terminal, but when
things go pear-shaped you <em>really</em> need to see what is on
the console. And you have to use the console to boot the
machine and do things like OS upgrades or installs.</p>
<p>You might try having a single console terminal and switching
from server to server as needed, either with a serial switch or
just by patching it into the required machine. Serial switches
are also hard to come by and not cheap, and may cause problems
with sending BREAK when they switch. And (if your computer room
is anything like ours) you never seem to have the right
combination of patch leads to connect to the machine you need
to, and even if the leads are there you can never work out
exactly which combination of DTE/DCE headshells goes with which
lead goes with which hardware. So you spend the first 10
minutes fooling around with breakout boxes and a box of leads,
all while the server is down and the users are screaming. Of
course this does not deal with the logging or remote access
requirements. And inevitably the console is not switched to the
machine you need so you lose all the console messages that might
tell you what is going on.</p>
<p>One popular solution is to use terminal server hardware.
Typically, the serial ports are connected to the various machine
consoles, and set up for "reverse telnet" access. This means a
user can telnet to a given IP/port and be connected to the
appropriate console. This can be very cost-effective, as
suitable old terminal servers can be picked up fairly cheaply
(assuming you do not have a couple lying around). And it is of
course network-accessible so suitable for remote access. But it
suffers from one major drawback: if the network is down, then
you have <em>no</em> access to any console, even if you are
standing right next to the machine. (This may be partially
alleviated by having a suitable terminal connected to one of the
terminal server ports and connecting from there, but the
terminal server software may not support that.) Also there is
no logging or replay of console messages. But with a bit of
work, and the addition of some software such as <a
href="http://www.conserver.com/">conserver</a> (described below),
this can be made to work pretty well.</p>
<p>A possibility suggested by Bron Gondwana is similar to the
above solution. If you use servers with multiple serial ports,
you can connect each spare serial port to the console port of
the "next" server, creating a ring of console connections (in
some sort of order). This can be made to work reasonably well
with the aid of the <a
href="http://www.conserver.com/">conserver</a> software, but can
be a bit confusing otherwise (i.e. remembering which port is
connected to which console). And you are stuck if you need to
use serial ports for other things (such as modems) or you have
machines without spare ports.</p>
<p>Or, if your budget exceeds your willingness to hack, you can
buy an off-the-shelf solution. These vary in price and
capability. See, for example, <a
href="http://www.lightwavecom.com/">Lightwave</a>, <a
href="http://www.perle.com/">Perle</a>, <a
href="http://www.avocent.com/">Avocent</a> or <a
href="http://www.blackbox.com/faxbacks/23000/23362.PDF">Black
Box</a>. These solutions can be quite expensive - typically
$USD100 - $USD400 per port.</p>
<h2>Our Solution</h2>
<p>In light of the above requirements, we chose a solution based
on a dedicated PC running Unix with a multiport serial card, and
some software designed to handle serial consoles.</p>
<p>It includes the following elements:</p>
<ul>
<li><p>A surplus PC. We used a Pentium 166, with a PCI bus,
2Gbyte hard disk and 64Mb of RAM. This is a massive overkill
for this task, and P-100, 500Mb, 32Mb would be more than
enough.</p></li>
<li><p>A PC Unix system. We used <a
href="../../../../index.html">FreeBSD 4.3</a> as that is used
for other tasks within our office.</p></li>
<li><p>A multi-port serial card. We chose the <a
href="http://www.stallion.com/html/products/easyio.html">EasyIO-PCI</a>
8-port card from <a href="http://www.stallion.com/">Stallion
Technologies</a>. This cost us about $AUD740, or under
$100/port, from <a href="http://www.ht.com.au/">Harris
Technologies</a> (which has lots of stuff but is by no means
the cheapest place in town - shop around and you might get it
a lot cheaper.) This card has a big DB80 connector on the
back, and a cable plugs into that which has a block with 8
RJ-45 sockets on it. (We chose the RJ-45 version as our
entire cable plant is RJ-45. This allows us to patch
connections from the required box to the console server
without any special cables.) This is the only thing we needed
to buy to make this all happen.</p></li>
<li><p>We build two servers, one for each computer room, with 8
ports in one and 16 ports (via two EasyIO-PCI cards) in the
other. If we needed more than 16 ports, then another of the
Stallion cards would be more cost-effective. We could
conceivably support 128 ports in each server (with 2
EasyConnect 8/64 host cards and 8 16 port RJ-45 modules) for
about $AUD12,000.</p></li>
<li><p>A modem for remote access to the console server host when
the network is down. We have not done this yet as the computer
room is next door, but when we put a server in Sydney we will
add the modem. The idea is that when the network is down, you
can dial up and log into the server machine and run the
console program locally. For security, we will probably leave
the modem powered off and ask the gopher in Sydney to turn on
the well-labelled button when we need it.</p></li>
<li><p>A program called <a
href="http://www.conserver.com/">conserver</a>. This program
does all the magic required to enable remote access to
consoles, and do the replaying and logging etc. It comes in
two parts: a server called <code>conserver</code> that runs as
a daemon and connects to the serial ports, handles logging
etc, and a client program called <code>console</code> that can
connect to the server, display console messages, send
keystrokes (and BREAKs) etc.</p></li>
</ul>
<p>This design covers all the major requirements except remote
power cycling:</p>
<ul>
<li><p>Remote access comes because the <code>console</code>
client program works across the network.</p></li>
<li><p>Logging is handled by the <code>conserver</code>
program.</p></li>
<li><p>If the network is down, then we can use the console on
the PC to run the <code>console</code> client locally. For
remote sites, we can add a modem for dial-in access to the the
server command line to run the client.</p></li>
<li><p>By patching the Solaris servers (see <a
href="#solaris">below</a>), we can avoid pranging the whole
computer room when the console server PC crashes (or the power
supply fails, or whatever).</p></li>
<li><p>We already have pager alerts from another system we have
installed, but the console server has all the required log
info so that could easily be implemented if we needed. And it
even has a modem for calling the pager company!</p></li>
<li><p>We do not currently support remote power cycling. Some
versions of the conserver program support this, but it does
require specialised serial-controlled power boards. We have
no immediate need for remote power cycling (we have a gopher
in each remote office who can do it by remote control) so this
is not a major problem, and we could add it easily should we
ever see the need and get the appropriate hardware.</p></li>
<li><p>This solution was very cheap. Total cost for the 9-port
server was $AUD750 for the IO card, as we re-used a surplus PC
and already owned the hardware for the special cables. If we
had to buy everything, then it would still only cost around
$AUD1500 for the 8-port server.</p></li>
</ul>
<h2>Setting Up The Server</h2>
<h4>Patching the Stallion driver</h4>
<p>The only hitch with setting up the server PC is getting the
device drivers for the Stallion card. FreeBSD has supported
Stallion ISA cards for many years, but unfortunately the driver
has not been actively maintained for some years, and does not
support the newer Stallion cards (such as the EasyIO-PCI card or
some of the other ISA cards with newer UART chips). I have put
together a <a href="stal-jumbo-patch">patch file</a> for FreeBSD
4.3 (based on the work of many other people) that will update
the system to support these newer cards. Hopefully, this patch
will be committed to the FreeBSD tree in time for the 4.4
release. If you are running FreeBSD 4 from prior to this time,
you will need to download and apply this patch (instructions are
included at the top of the patch file). If you are running
FreeBSD 3, then you had probably best upgrade.</p>
<h4>Configuring a new kernel</h4>
<p>The Stallion driver is not included in the default
<samp>GENERIC</samp> kernel, so you will need to create a kernel
config file with the appropriate entries. See the
<samp>stl(4)</samp> manual page and the appropriate section of the
<a
href="../../../../doc/en_US.ISO8859-1/books/handbook/kernelconfig.html">FreeBSD
Handbook</a>.</p>
<h4>Building World</h4>
<p>After applying the patch you will certainly need to update the
whole system, including all the user programs and the kernel.
See the documentation in <samp>/usr/src/UPDATING</samp> and the
<a href="http://www.FreeBSD.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html">FreeBSD
Handbook</a>.</p>
<h4>Making The Devices</h4>
<p>You will need to make the device nodes for the Stallion card
(which are not made by default). A new version of
<samp>/dev/MAKEDEV</samp> with Stallion support will have been
created by the <samp>mergemaster</samp> run during the above
procedure. If you have a Stallion card with more than 8 ports,
then you will need to edit <samp>/dev/MAKEDEV</samp> and change
the definition of <samp>maxport</samp> at about line 250. By
default, <samp>MAKEDEV</samp> only makes device nodes for 8
ports to keep the size of the <samp>/dev</samp> directory down.</p>
<p>Run a command like</p>
<pre>
cd /dev/ && sh MAKEDEV cuaE0
</pre>
<p>to create dial-out devices for the first Stallion card. See
the comments in <samp>MAKEDEV</samp> and the <samp>stl(4)</samp>
manual page for more details.</p>
<h4>Compiling conserver</h4>
<p>(See the section on Conserver versions <a
href="#conserver-versions">below</a>; the version I use is not
the one available in the FreeBSD ports collection.)</p>
<p>There are two ways to install conserver. You can either
compile from the source or use the FreeBSD ports framework.</p>
<h5>Using the port framework</h5>
<p>Using the ports framework is a bit cleaner, as the package
system can then keep track of installed software and cleanly
delete them when not being used. Download the port framework
from <a href="conserver-port.tar.gz">here</a>, unpack it to
create a directory called <code>conserver-port</code>. CD to
that directory and (as <code>root</code>) type</p>
<pre>
make DEFAULTHOST=<em>consolehost</em> install
</pre>
<p>where <em>consolehost</em> is the name of the machine running
the console server. Specifying this when the binary is compiled
will avoid having to either specify it each time the program is
run on remote hosts or having to maintain a
<code>conserver.cf</code> file on every host. This command will
fetch, patch, configure, compile and install the conserver
application.</p>
<p>You can then run <samp>make package</samp> to create a binary
package that can be installed on all the other FreeBSD hosts
with the <samp>pkg_add</samp> command. For extra style points,
you can make a two versions of the package: one for the console
server machine without a <samp>DEFAULTHOST</samp> argument, and
one for all the other hosts with a <samp>DEFAULTHOST</samp>
argument. This will mean the console client program on the
console server machine will default to <samp>localhost</samp>,
which will work in the absence of name servers when the network
is busted, and also allow "trusted" (i.e. no password required)
connections via the localhost IP address for users logged into
the console server machine (either via the console screen or the
emergency backup modem). The version for the other machines
with a <samp>DEFAULTHOST</samp> argument means users can just
use the <samp>console</samp> client without specifying a
hostname every time, and without needing to configure the
<samp>conserver.cf</samp> file on every machine.</p>
<h5>From the source tarball</h5>
<p>If you prefer, you can download conserver and compile it
yourself. You might need to do this if you want to install the
console client on non-FreeBSD systems. We run the client on our
Solaris hosts and it inter-operates with the FreeBSD-hosted
server with no problems. This allows anyone in the whole
company (many of whom have PCs and no FreeBSD host access on
their desk) to access the console server.</p>
<p>Download the file from the <a
href="ftp://ftp.conserver.com/conserver/conserver-7.0.2.tar.gz">conserver.com
FTP site</a>. Extract it into a handy directory then configure
it by running</p>
<pre>
./configure --with-master=<em>consoleserver</em> --with-port=782
</pre>
<p>The <code>--with-master</code> argument avoids having to
specify the master server every time the client is run remotely
(or keeping up-to-date config files on all remote hosts). The
<code>--with-port</code> argument avoids having to update
<code>/etc/services</code> on every machine.</p>
<p>Then type <code>make</code> and, as root, <code>make
install</code>.</p>
<h4>Configuring conserver</h4>
<p>The conserver program is configured via a file called
<code>conserver.cf</code>. This file usually lives in
<code>/usr/local/etc</code> and is documented in the <a
href="conserver.cf.html"><code>conserver.cf(5)</code></a> man
page.</p>
<p>Our config file looks like this:</p>
<pre>
LOGDIR=/var/log/consoles
gallows:/dev/cuaE0:9600p:&:
roo:/dev/cuaE1:9600p:&:
kanga:/dev/cuaE2:9600p:&:
%%
allow: itga.com.au
trusted: 127.0.0.1 buzz
</pre>
<p>The first line means all the console log files by default go
into the directory <code>/var/log/consoles/</code>. The
<samp>&</samp> in each line says the log file for that machine
will be <samp>/var/log/consoles/</samp><em>machine</em>.
<p>The next three lines show three machines to which we need to
connect. We use the cuaE<em>x</em> devices rather than the
ttyE<em>x</em> devices because console ports typically do not
show carrier. This means that opening ttyE<em>x</em> would hang
and <code>conserver</code> would never connect. Using the
cuaE<em>x</em> device avoids this problem. Another solution
would be to use the ttyE<em>x</em> devices and enable <em>soft
carrier</em> on these ports, perhaps by setting this using the
ttyiE<em>x</em> device in the <code>/etc/rc.serial</code> file.
See the comments in this file for more details. Also see the
<code>sio(4)</code> manual page for information on the
initial-state and locked-state devices. (The Stallion driver
also supports these conventions). And see the
<code>stty(1)</code> manual page for details on setting device
modes.</p>
<p>The last section shows that any user logged into the server
machine has passwordless access to all consoles. We do this
because there are no user accounts on this machine and it is
safely isolated from the wide world behind our firewall. The
allow line allows anyone on a machine inside our organisation to
access the console server if they provide their password, which
is recorded in the <code>conserver.passwd</code> file (see next
section).</p>
<h4>Setting conserver passwords</h4>
<p>The <code>conserver.passwd</code> file contains the encrypted
version of the password that each user. The file is documented
in the <a
href="conserver.cf.html"><code>conserver.cf(5)</code></a> man
page.</p>
<p>The only tricky bit is loading the file with encoded passwords.
It appeared in FreeBSD that was is no obvious way to generate an
encrypted password for inclusion in another file (but see
below). So I put together a quick hack perl script to do
this:</p>
<pre><![ CDATA [/*
@rands = ();
foreach (0..4) {
push(@rands, rand 64);
}
$salt = join '', ('.', '/', 0..9, 'A'..'Z', 'a'..'z')[@rands];
$salt = '$1$' . $salt . '$';
print 'Enter password: ';
`stty -echo`;
$cleartext = <>;
`stty echo`;
chop($cleartext);
print crypt($cleartext, $salt), "\n";
]]></pre>
<p>(Note that this uses the FreeBSD MD5-style encrypted passwords.
Running this on other Unix variants, or on FreeBSD with DES
passwords, will likely need a different style of salt.)</p>
<p>Kris Kennaway has since pointed out you can get the same effect
using the <samp>openssl passwd</samp> command:</p>
<pre>
$ openssl passwd -1
Password: &lt;password&gt;
$1$VTd27V2G$eFu23iHpLvCBM5nQtNlKj/
$
</pre>
<h4>Starting conserver at system boot time</h4>
<p>There are two ways this can be done. Firstly, you could start
up conserver from init by including an entry in /etc/ttys that
looks a bit like this:</p>
<pre>
cuaE0 "/usr/local/sbin/conserver" unknown on insecure
</pre>
<p>This has two advantages: init will restart the master console
server if it ever crashes for any reason (but we have not
noticed any crashes so far), and it arranges for STDOUT of the
conserver process to be directed to the named tty (in this case
cuaE0). This is useful because you can plug a terminal into
this port, and the conserver program will show all console
output not otherwise captured by a client console connection.
This is useful as a general monitoring tool to see if anything
is going on. We set this terminal up in the computer room but
visible from the main office. It is a very handy feature. The
downside of running conserver from the ttys file is that it
cannot run in daemon mode (else init would continually restart
it). This means conserver will not write a PID file, which makes
it hard to rotate the log files.</p>
<p>So we start conserver from an rc.d script. If you installed
conserver via the port, there will be a
<code>conserver.sh.sample</code> file installed in
<code>/usr/local/etc/rc.d</code>. Copy and/or rename this to
<code>conserver.sh</code> to enable conserver to start at boot
time.</p>
<p>In fact we use a modified version of this script which also
connects conserver to a terminal via a tty device so we can
monitor unwatched console output. Our conserver.sh script looks
like this:</p>
<pre>
#!/bin/sh
#
# Startup for conserver
#
PATH=/usr/bin:/usr/local/bin
case "$1" in
'start')
TTY=/dev/cuaE7
conserver -d > $TTY
# get NL->CR+NL mapping so msgs look right
stty < /dev/cuaE7 opost onlcr
echo -n ' conserver'
;;
'stop')
kill `cat /var/run/conserver.pid` && echo -n ' conserver'
;;
*)
echo "Usage: $0 { start | stop }"
;;
esac
exit 0
</pre>
<p>(Note the use of cuaE0 device and the need to set TTY modes for
proper NL->CR handling).</p>
<h4>Keeping the log files trimmed</h4>
<p>FreeBSD has a program called <code>newsyslog</code> that will
automatically handle log file trimming. Just add some lines to
the configuration file <code>/etc/newsyslog.conf</code> for the
console logs:</p>
<pre>
#
# The log files from conserver
/var/log/consoles/gallows 644 10 1000 * Z /var/run/conserver.pid
/var/log/consoles/kanga 644 10 1000 * Z /var/run/conserver.pid
/var/log/consoles/roo 644 10 1000 * Z /var/run/conserver.pid
</pre>
<p>This tells newsyslog (which is run from cron every hour on the
hour) that the console log files should be archived and
compressed once they reach 1Mb, that we should keep 10 of them,
and that to signal the server program you send a SIGHUP to the
process whose PID is in the conserver.pid file. This is the
master server, and it will arrange to signal all the child
processes. Yes, this will send a HUP to all clients whenever a
single log file needs rotating, but that is quite cheap. See
the newsysylog(8) manual page for details.</p>
<h2>Cabling</h2>
<p>This is always the hardest part of this kind of problem!! We
had only a dozen or so cables/headshells to build, and we
already had a collection of the appropriate crimping tools and
hardware, so we did it ourselves. But if you are not set up for
this, or you have a large number of cables to make, then you
might consider getting some cables custom made. Look in the
yellow pages, there are a surprising number of places that do
this! Getting custom-made cabling is good, and you can get much
more professional results, but can be expensive. For example,
the RJ-45 to DB-25 adapter kits described below are about $10
each; custom-made headshells are about twice that (and take a
couple of weeks to arrive). Similarly, crimping custom RJ-45 to
RJ-45 leads is quite cheap (say, $5 each) but it takes a fair
amount of time. Custom made RJ-45 socket to RJ-45 plug
converters cost about $25 each.</p>
<p>We have settled on RJ-45 Cat-V cabling for all our office and
computer room cabling needs. This included patching between
racks in the computer room. For serial connections, we use
patchable headshells that have RJ-45 sockets on the back. This
allows us to patch whatever RJ-45 - DB-25 connections we need.</p>
<p>Which is just as well, because there are many incompatible ways
to represent serial connections on the RJ-45 plug. So the
cabling has to be very careful to use the right mapping.</p>
<h4>RJ-45 colors</h4>
<p>RJ-45 cables and plugs have 8 pins/conductors. These are used
as 4 matched pairs. There are a couple of conventions about how
the pairs are mapped onto pins, but 100baseT uses the most
common (known as EIA 586B). There are three common color-coding
conventions for the individual conductors in RJ-45 cables. They
are:</p>
<table>
<tr class="alt">
<th>Pin</th>
<th>Scheme 1</th>
<th>Scheme 2 (EIA 568B)</th>
<th>Scheme 3 (EIA 568A)</th>
<th>Pair</th>
</tr>
<tr>
<td>1</td>
<td>Blue</td>
<td>White+Green</td>
<td>White+Orange</td>
<td>2+</td>
</tr>
<tr class="alt">
<td>2</td>
<td>Orange</td>
<td>Green</td>
<td>Orange</td>
<td>2-</td>
</tr>
<tr>
<td>3</td>
<td>Black</td>
<td>White+Orange</td>
<td>White+Green</td>
<td>3+</td>
</tr>
<tr class="alt">
<td>4</td>
<td>Red</td>
<td>Blue</td>
<td>Blue</td>
<td>1+</td>
</tr>
<tr>
<td>5</td>
<td>Green</td>
<td>White+Blue</td>
<td>White+Blue</td>
<td>1-</td>
</tr>
<tr class="alt">
<td>6</td>
<td>Yellow</td>
<td>Orange</td>
<td>Green</td>
<td>3-</td>
</tr>
<tr>
<td>7</td>
<td>Brown</td>
<td>White+Brown</td>
<td>White+Brown</td>
<td>4+</td>
</tr>
<tr class="alt">
<td>8</td>
<td>White<br>or Grey</td>
<td>Brown</td>
<td>Brown</td>
<td>4-</td>
</tr>
</table>
<p>Note EIA 468A and EIA 568B are very similar, simply swapping
the colors assigned to pair 2 and pair 3.</p>
<p>See for example the <a
href="http://www.cabletron.com/support/techtips/tk0231-9.html">Cabletron
Tech Support Site</a> for more details.</p>
<p>The pins in the RJ-45 plug are numbered from 1 to 8. Holding a
patch lead with the cable pointing down and the clip away from
you, pin 1 is at the left. Or, looking into an RJ-45 socket
with the clip to the top, pin 1 is on the right. The following
illustration (shamelessly lifted from the Cabletron web site
above) shows it pretty well:</p>
<img alt="RJ45" src="tk0231-9-1.png">
<p>We have four classes of equipment to deal with in our setup:</p>
<dl>
<dt>Sun Servers
<dd>
<p>Sun servers operate as DTE (i.e. send data on TxD and read RxD,
and assert DTR) with a female DB-25 socket on board. So we need
to create a headshell for the Stallion that operates as DCE and
has a male DB-25 plug (i.e. acts as a <em>null modem</em> cable
as well as converts from RJ-45 to DB-25). We use headshells
that have an RJ-45 socket in them and 8 short flyleads with
DB-25 pins on the end. These pins can be inserted into the
DB-25 plug as required. This allows us to create a custom
RJ-45-DB-25 mapping. We used a couple of different sorts,
including the <a href="http://www.molexpn.com.au/">MOD-TAP</a>
part no. <a
href="http://www.molexpn.com.au/products/index.nsx/1/7/0/0/id=340">06-9888-999-00</a>
and the <a
href="http://www.blackbox.com/faxbacks/12000/12654.PDF">FA730
series</a> from <a href="http://www.blackboxoz.com.au/">Black
Box</a>.</p>
<p>On our version of the headshells, these flyleads had the
following colours (from Pin 1-8): Blue, Orange, Black, Red,
Green, Yellow, Brown, White. (Looking into an RJ-45 socket,
with the clip towards the top, pin 1 is on the right.) This is
how they are connected to the DB-25 socket:</p>
<table>
<tr class="alt">
<th>Stallion RJ-45 Pin</th>
<th>Colour</th>
<th>Signal</th>
<th>Sun DB-25 Male Pin</th>
<th>RS232 Signal</th>
</tr>
<tr align=center>
<td>1</td>
<td>Blue</td>
<td>DCD</td>
<td>20</td>
<td>DTR</td>
</tr>
<tr align=center class="alt">
<td>2</td>
<td>Orange</td>
<td>RTS</td>
<td>5</td>
<td>CTS</td>
</tr>
<tr align=center>
<td>3</td>
<td>Black</td>
<td>Chassis Gnd</td>
<td>1</td>
<td>Chassis Gnd</td>
</tr>
<tr align=center class="alt">
<td>4</td>
<td>Red</td>
<td>TxD</td>
<td>3</td>
<td>RxD</td>
</tr>
<tr align=center>
<td>5</td>
<td>Green</td>
<td>RxD</td>
<td>2</td>
<td>TxD</td>
</tr>
<tr align=center class="alt">
<td>6</td>
<td>Yellow</td>
<td>Signal Gnd</td>
<td>7</td>
<td>Signal Gnd</td>
</tr>
<tr align=center>
<td>7</td>
<td>Brown</td>
<td>CTS</td>
<td>4</td>
<td>RTS</td>
</tr>
<tr align=center class="alt">
<td>8</td>
<td>White</td>
<td>RTS</td>
<td>8</td>
<td>DCD</td>
</tr>
</table>
<p>Note that colours may be different for your cables/headshells.
In particular, pin 8 may be grey instead of white.</p>
<p>Remember to label the headshell <em>clearly</em>, in a way that
will not fade/fall off/rub off with time!</p>
<dt>Cisco 16xx/26xx/36xx Routers
<dd>
<p>I think that all Cisco gear that has RJ-45 console ports and
runs IOS will have the same cable requirements. But best to
check first. We have tried this on 1600s and 2600s only.</p>
<p>Both the Stallion card and the 2600 have RJ-45 connections, but
of course they are not compatible. So you need to crimp up a
special RJ-45-RJ-45 cable. And this cable must be plugged in
the right way round! We use normal RJ-45 flyleads from the
router to the patch panel, then the special flylead from the
patch panel to the Stallion card.</p>
<p>We built two special Stallion-Cisco leads by cutting in half a
2m flylead and crimping an RJ-45 with the appropriate pinouts to
each free end. The original connector will be the Cisco end of
the cable, the new crimped connector will be the Stallion end.
Holding the RJ-45 connector on the flylead with the cable
pointing down and the clip pointing away, this is the order of
the colours of the cables in our flylead (pins 1-8, from L to
R): white/green, green, white/orange, blue, white/blue, orange,
white/brown, brown. For the Stallion end, trim and discard the
brown/white+brown and green/white+green pairs. Then holding the
RJ-45 plug in the same manner (cable down, clip away), the
connections should be (from L to R): None, None, Blue, Orange,
White/Orange, White/Blue, None, None, as shown:</p>
<table>
<tr class="alt"><th>Cisco RJ-45 Pin<th>Colour<th>Cisco Signal<th>Stallion RJ-45
Pin<th>Stallion Signal</tr>
<tr align=center><td>1<td>White/Green<td>RTS<td>N/C<td>&nbsp;</tr>
<tr align=center class="alt"><td>2<td>Green<td>DTR<td>N/C<td>&nbsp;</tr>
<tr align=center><td>3<td>White/Orange<td>TxD<td>5<td>RxD</tr>
<tr align=center class="alt"><td>4<td>Blue<td>Gnd<td>3<td>Gnd</tr>
<tr align=center><td>5<td>White/Blue<td>Gnd<td>6<td>Gnd</tr>
<tr align=center class="alt"><td>6<td>Orange<td>RxD<td>4<td>TxD</tr>
<tr align=center><td>7<td>White/Brown<td>DSR<td>N/C<td>&nbsp;</tr>
<tr align=center class="alt"><td>8<td>Brown<td>CTS<td>N/C<td>&nbsp;</tr>
</table>
<p>Note again that colours may be different for your cables/headshells.</p>
<p>Carefully label the cable, and each end of the cable, and test
it. If it does not work, testing is <em>really</em> hard as they
do not make RJ-45 serial line testers!</p>
<p>Let me state this more stongly: Be <em>very</em> <b>VERY</b>
sure that you label this cable in a way that is easily,
instantly and permanently recognisable as a special cable and
not easily confused with normal drop cables. Some suggestions
(from Hugh Irvine):</p>
<ul>
<li> Make them out of different coloured cable
<li> For marking the ends, clear heat-shrink tubing slipped over
printed labels *before* putting on the connectors is the best way I have
seen for marking what they are.
<li>You can also use Panduit or similar tags that
you put on with nylon tie straps, but I find the ink wears off the
tags.
</ul>
<p>
<dt>Cisco Catalyst switches
<dd>
Astoundingly, the pinout on the console ports of the Catalysts is
actually <em>different</em> to the pinout used on the 26xx-series
Cisco hardware. I think the way to tell which is which is by
considering the operating software. If it uses IOS, then the previous
pinout is required. If it uses the switch software, then this pinout
is required.
<p>Fortunately, while the pinouts are different, the Catalyst
pinout is simply a mirror image of the pinout for the 2600.
Even more fortunately, the Ciscos (both Catalysts and 2600s)
seem to ship with a special "rollover" cable, which is exactly
what is required in this case. We use the rollover cable from
the Catalysts to the patch panel, then the same cable as above
for the 2600s from the patch panel to the Stallion card, and it
all works just fine.</p>
<p>This rollover cable is an RJ-45-RJ-45 cable and is intended to
be used with the shipped (hardwired) RJ-45 - DB-25 and RJ-45 -
DB-9 headshells for console connections. Ours are 2m long,
either light blue or black, and are quite flat. Attempts to use
them for 100baseT ethernet will fail miserably! You can tell it
is a rollover cable by holding both ends with the cable pointing
down and the clip pointing away from you. Check the colour of
the leads in each pin in the two connectors, they should be
mirror images. (In our case, one goes
grey-orange-black-red-green-yellow-blue-brown, the other
brown-blue-yellow-green-red-black-orange-grey). This is a
rollover cable.</p>
<p>If you do not have a rollover cable present, then you can use
the same cable as for the 26xx except plug it in the other way
around (i.e. original 8-pin plug goes into the Stallion, the new
crimped plug with only 4 active wires goes into the Catalyst).</p>
<p>
<dt>FreeBSD Servers (or any other i386 PC systems using a serial console)
<dd>
<p>We run FreeBSD 4 on a couple of i386 PCs for various peripheral
uses. FreeBSD usually uses a screen and keyboard for the
console, but can be configured to use a serial port (usually the
first serial port known as COM1 in DOS/Windows or ttyd0 in
Unix).</p>
<p>The cabling for these servers depends on the PC harware. If
the PC has DB-25 female socket on board (as most older PCs do),
then the same headshell as works for the Sun server above will
work fine. If the PC has DB-9 male plug on board (as more
recent PCs tend to do), then there are two choices. Either use
a DB-9 to DB-25 converter (this is not recommended as it can
lead to unreliable connections over the long term as the adapter
is bumped/works loose), or build an RJ-45 to DB-9 cable as
follows:</p>
<table>
<tr class="alt"><th>Stallion RJ-45 Pin<th>Colour<th>Signal<th>PC DB-9 Female
Pin<th>RS232 Signal</tr>
<tr align=center><td>1<td>Blue<td>DCD<td>4<td>DTR</tr>
<tr align=center class="alt"><td>2<td>Orange<td>RTS<td>8<td>CTS</tr>
<tr align=center><td>3<td>Black<td>Chassis Gnd<td>N/C<td>&nbsp;</tr>
<tr align=center class="alt"><td>4<td>Red<td>TxD<td>2<td>RxD</tr>
<tr align=center><td>5<td>Green<td>RxD<td>3<td>TxD</tr>
<tr align=center class="alt"><td>6<td>Yellow<td>Signal
Gnd<td>5<td>Signal Gnd</tr>
<tr align=center><td>7<td>Brown<td>CTS<td>7<td>RTS</tr>
<tr align=center class="alt"><td>8<td>White<td>RTS<td>1<td>DCD</tr>
</table>
See <a href="#freebsd">below</a> for tips on configuring FreeBSD to
use a serial console.
</dl>
<h2><a name="solaris">On Sun Systems And Break</a></h2>
<p>Anyone who has turned off a terminal used as a console for a
Sun system will know what happens and why this is a problem.
Sun hardware recognises a serial BREAK as a command to halt the
OS and return to the ROM monitor prompt. A serial BREAK is an
out-of-band signal on an RS-232 serial port that involves making
the TX DATA line active (i.e. pulled down to less than -5v) for
more than two whole character times (or about 2ms on a 9600bps
line). Alas, this BREAK signal is all to easily generated by
serial hardware during power-on or power-off. And the Stallion
card does, in fact, generate breaks when the power to the PC
fails. Unless fixed, this problem would mean that every Sun box
connected to the console server would be halted whenever the
power failed (due to dead power supplies, or fat-fingered
operators unplugging it, or whatever). This is clearly not an
acceptable situation.</p>
<p>Fortunately, Sun have come up with a set of fixes for this.
For Solaris 2.6 and later (and perhaps earlier, but who was
crazy enough to run 2.5??), the <code>kbd(1)</code> command can
be used to disable the ROM-on-BREAK behaviour. This is a good
start, but leaves you out of luck in the situation where a break
is needed to get into a broken machine.</p>
<p>Starting with Solaris 8, the <code>kbd</code> command can also
be used to enable an alternate break sequence using the
<samp>kbd -a alternate</samp> command. When this is set, the
key sequence &lt;Return&gt;&lt;Tilda&gt;&lt;control-B&gt;
(within 5 seconds) will drop to the ROM. You can enable this
permanently by editing the <code>/etc/default/kbd</code> file;
see the <code>kbd(1)</code> manual page. Note that this alternate
break sequence is only active once the kernel has started
running multiuser and processed the default file. While the ROM
is active (during power-on and during the boot process) and
while running single-user, you still need to use a BREAK to get
to the ROM prompt. The console client can cause the server to
send a BREAK using the escape sequence "^Ecl1" (i.e. escape, c,
ell, one).</p>
<p>If you have a Sun software support contract, there are patches
available for Solaris 2.6 and 2.7 that add the"alternate break"
capability integrated into Solaris 2.8. Solaris 2.6 requires
patch 105924-10 or higher. Solaris 2.7 requires patch 107589-02
or higher.</p>
<p>We have added this patch to all our Solaris 2.6 servers, and
added it (and the entry in the /etc/default/kbd file) to our
jumpstart configuration so it will automatically be added to
every new install.</p>
<p>We have confirmed by direct testing that neither the Cisco
16xx, 26xx, or Catalyst hardware suffers from the BREAK sent
when the Stallion card loses power.</p>
<h2><a name="freebsd">Using a Serial Console on FreeBSD</a></h2>
<p>The procedure for doing this is described in detail in the <a
href="../../../../doc/en_US.ISO8859-1/books/handbook/serialconsole-setup.html">FreeBSD
Handbook</a>. This is a quick summary.</p>
<h4>Check the kernel configuration</h4>
<p>Check that the kernel configuration file has <samp>flags
0x10</samp> in the config line for the <samp>sio0</samp> device.
This signals this device (known as <samp>COM1</samp> in
DOS/Windows or <samp>/dev/ttyd0</samp> in FreeBSD) can be used
as a console. This flag is set on the <samp>GENERIC</samp> and
<samp>LINT</samp> sample configs, so is likely to be set in your
kernel.</p>
<h4>Create the <samp>/boot.conf</samp> file</h4>
<p>This file should be created containing a single line containing
just "-h". This tells the FreeBSD boot blocks to use the serial
console.</p>
<h4>Edit <samp>/etc/ttys</samp></h4>
<p>Edit this file and make the following changes.</p>
<p>If you are not going to have any keyboard/video screen on this
server at all, you should find all the lines for
<samp>ttyv0</samp> through <samp>ttyv8</samp>:</p>
<pre>
ttyv1 "/usr/libexec/getty Pc" cons25 on secure
</pre>
<p>Change the <samp>on</samp> to <samp>off</samp>. This will stop
login screens being run on the useless video consoles.</p>
<p>Find the line containing <samp>ttyd0</samp>. Change it
from</p>
<pre>
ttyd0 "/usr/libexec/getty std.9600" dialup off secure
</pre>
<p>to</p>
<pre>
ttyd0 "/usr/libexec/getty std.9600" vt100 on secure
</pre>
<p>(replacing <samp>vt100</samp> with the term type of your
console. <samp>xterms</samp> might be a good choice). This
allows you to log in to the console port once the system is
running multi-user.</p>
<p>Reboot and away you go!</p>
<h2>Security Implications</h2>
<p>The client-server protocol for <code>conserver</code> requires
the user of the <code>console</code> client to enter a password.
This password is passed across the net in <em>cleartext!!!</em>
This means <code>conserver</code> is not really suitable for use
across untrusted networks (such as the Internet). Use of
conserver-only passwords (in the <code>conserver.passwd</code>
file) slightly mitigate this problem, but anyone sniffing a
conserver connection can easily get console access, and from
there prang your machine using the console break sequence. For
operating across the Internet, use something secure like
<em>SSH</em> to log into to the server machine, and run the
console client there.</p>
<h2><a name="conserver-versions">On Conserver Versions</a></h2>
<p>The <code>conserver</code> program has fractured into a number
of versions. The home page referenced above seems to be the
latest and most featureful version around, and carries a version
number of "7.0.2". This is maintained by Bryan Stansell (<a
href="mailto:bryan@conserver.com">bryan@conserver.com</a>), who
has brought together the work of many people (listed on his
webpage).</p>
<p>The FreeBSD ports collection contains a port for version 8.5 of
conserver. This seems to be older and less featureful than the
7.0.2 version (in particular, it does not support consoles
connected to terminal server ports and does not support a
<code>conserver.passwd</code> file), and is written in a fairly
idiosyncratic manner (using a preprocessor to generate C code).
Version 8.5 is maintained by Kevin S Braunsdorf (<a
href="mailto:ksb+conserver@sa.fedex.com">ksb+conserver@sa.fedex.com</a>)
who did most of the original work on conserver, and whose work
Bryan Stansell is building on. The 8.5 version does support one
feature not in the 7.0.2 version (controlling power to remote
machines via a specific serial-interfaced power controller
hardware).</p>
<p>As shipped with FreeBSD, the 8.5 version does not authenticate
against FreeBSD servers using MD5 passwords. There is a patch
in the FreeBSD GNATS system (<a
href="http://www.FreeBSD.org/cgi/query-pr.cgi?pr=28432">ports/28432</a>)
that fixes this problem. Kevin has indicated this patch will be
included in future versions of conserver from him. There are a
couple of other problems with the 8.5 version as well.</p>
<ul>
<li><p>It is hard to rotate log files, as <code>conserver</code>
does not create a PID file and killing the parent conserver
process does not kill the children (which have the logfiles
open).</p></li>
<li><p>Compiling it (on anything other than FreeBSD) is a pain
because it depends on a couple of other large programs to
compile. This is because it is not written in C but in a
macro language that produces C code, and so depends on these
macro processors. The intent of this macro language is to
hide system dependencies. The 7.0.2 version uses GNU
configure for the same task, and compiling this version on
Solaris is a breeze. (Compiling 8.5 is easier on FreeBSD as
the ports framework takes care of all this for you).</p></li>
<li><p>After compiling on Solaris, the <code>console</code>
program would not authenticate against the
<code>conserver</code> program on FreeBSD, because it used
<code>getpass()</code> on Solaris that silently truncates
passwords to 8 characters. FreeBSD uses MD5 passwords that
can be (and in our case, are) longer than 8 characters.</p>
<li><p>There was some confusion about the role of "groups" in
<code>conserver</code>. When replaying log messages in 8.5,
it seemed to replay messages from all consoles in the group,
whereas 7.0.2 seems to do the obvious thing and keep output
from individual consoles separate. (This may well have just
been boneheaded setup mistakes on my part however.) The
upshot was that the 8.5 version needed a new group (and
process) per console, whereas the 7.0.2 version will happily
support multiple ports per process.</p></li>
</ul>
<H2>Links</h2>
<p><a href="http://www.conserver.com/">Homepage</a> for the latest
version of conserver.</p>
<p><a
href="ftp://ftp.conserver.com/conserver/conserver-7.0.2.tar.gz">FTP</a>
the conserver 7.0.2 source tarball.</p>
<p><a href="http://www.stallion.com/">Stallion Technologies</a> homepage.</p>
<p>David Harris' <a
href="http://www.conserver.com/consoles/msock.html">Minor Scroll
Of Console Knowledge</a> contains heaps of good information on
serial consoles and serial comms in general.</p>
<p>His <a href="http://www.conserver.com/consoles/">Greater Scroll
of Console Knowledge</a> contains evern more specific info on
connecting various devices to various other devices. Oh the
joys of standards!</p>
<p>Doug Hughes has <a
href="http://www.eng.auburn.edu/users/doug/console.html">a
similar console server</a> based on the <samp>screen</samp>
program and an old SunOs host.</p>
<p>The <a href="http://www.realweasel.com/">Real Weasel</a> company
makes a ISA or PCI video card that looks like a PC video card
but actually talks to a serial port. This can be used to
implement serial consoles on PC hardware for operating systems
that can not be forced to use serial console ports early
enough.</p>
<h2>Files</h2>
<a href="stal-jumbo-patch">Stallion Jumbo Patch for FreeBSD 4
(unzipped, 400k)</a><br> The same <a
href="stal-jumbo-patch.gz">Stallion Jumbo Patch for FreeBSD 4
(gzipped, 140k)</a><br>
A FreeBSD <a href="conserver-port.tar.gz">port framework</a> for
conserver 7.0.2.<br> The <a href="conserver-port.shar">Port
framework as a SHAR file</a><br>
<h2>Man Pages</h2>
<a href="console.html">console(8)</a><br>
<a href="conserver.html">conserver(8)</a><br>
<a href="conserver.cf.html">conserver.cf(5)</a><br>
<h2>Change Log</h2>
<dl>
<dt>10 July, 2001
<dd>Initial version announced on FreeBSD and Sage-AU mailing lists.
<dt>12 July, 2001
<dd>
<ul>
<li>Notes on some commercial console servers from Doug Schache.
<li>Grrr. Fixed the pinout for the Cisco - Stallion RJ-45 cable.
The cable would have worked but was backwards. The new table now
reflects the description.
<li>Noted that Catalyst cable is a 26xx cable plugged in the other
way around.
<li>Added notes about RJ-45 pin numbering and color coding
conventions
<li>Added notes about serial console on FreeBSD/i386 platforms.
<li>Now in RCS.
</ul>
<dt>13 July, 2001
<dd>
<ul>
<li>Notes on cable marking from Hugh Irvine
<li>Notes on custom-made cables, source for headshells
</ul>
<dt>18 July, 2001
<dd>
<ul>
<li>Links to Black Box adapters
<li>Fix dates. It's July, already!
<li>Pointers to commercial solutions and serial-port-on-other-server setup
<li>Pointer to "openssl passwd" command
</ul>
</dl>
<hr>
<address><a href="mailto:gnb@itga.com.au">Gregory Bond</a></address>
<!-- Created: Wed Jul 4 12:35:02 EST 2001 -->
<!-- hhmts start -->
Last modified: Thu Jul 19 10:19:28 EST 2001
<!-- hhmts end -->
<br>
$Id: article.sgml,v 1.8 2002-07-11 19:07:48 trhodes Exp $
</body>
</html>