...making Linux just a little more fun!

December 2005 (#121):


The Front Page

By Heather Stern

A
Little Xen in my garden

[BIO]

XenSource is the company that commercially supports the Xen hypervisor environment and support tools for using it. The software uses paravirtualization for high performance, running several guest domains on one parent OS.

I would describe Zen, but if it can be described successfully, would it really need a description anymore? :) The gate to an open mind and heart is inside.

This really is the inside gate of my garden. Guests in my domain are welcome to coffee anytime!


Heather is Linux Gazette's Technical Editor and The Answer Gang's Editor Gal.


[BIO] Heather got started in computing before she quite got started learning English. By 8 she was a happy programmer, by 15 the system administrator for the home... Dad had finally broken down and gotten one of those personal computers, only to find it needed regular care and feeding like any other pet. Except it wasn't a Pet: it was one of those brands we find most everywhere today...

Heather is a hardware agnostic, but has spent more hours as a tech in Windows related tech support than most people have spent with their computers. (Got the pin, got the Jacket, got about a zillion T-shirts.) When she discovered Linux in 1993, it wasn't long before the home systems ran Linux regardless of what was in use at work.

By 1995 she was training others in using Linux - and in charge of all the "strange systems" at a (then) 90 million dollar company. Moving onwards, it's safe to say, Linux has been an excellent companion and breadwinner... She took over the HTML editing for "The Answer Guy" in issue 28, and has been slowly improving the preprocessing scripts she uses ever since.

Here's an autobiographical filksong she wrote called The Programmer's Daughter.

Copyright © 2005, Heather Stern. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

Firewall logging to MySQL - the quick and easy way

By Anonymous

Security is a journey, not a destination. One good step along the way is to review and analyze your firewall logs and syslog messages on a regular basis. Unfortunately, the plain text logs produced by syslog are not in a form that is easily analyzed. Also, unless you are using syslog-ng, your firewall logs are probably scattered all over the various system message log files.

This article will show you how to move your firewall logs from syslog text files to a MySQL database in 10 minutes or so. The following examples were carried out on a SuSE 10.0 system but you can easily adapt them for other distributions.

1. Verify kernel settings

You can skip this step if you are using the default SuSE 10.0 kernel. The stock kernels that come with most distributions should be fine, but you will need to make sure you have your kernel compiled with the CONFIG_NETFILTER, CONFIG_IP_NF_IPTABLES, CONFIG_IP_NF_FILTER, and CONFIG_IP_NF_TARGET_ULOG options. Most firewalls will also need CONFIG_IP_NF_CONNTRACK, CONFIG_IP_NF_FTP, and CONFIG_IP_NF_IRC.

If you have a file called /proc/config.gz, it means your kernel was compiled with the IKCONFIG option. /proc/config.gz is the compressed version of the .config file that was used to generate that kernel, so you can check if you have the necessary options for netfilter and ulog with this command:
gunzip -c /proc/config.gz | grep -E 'CONFIG_(NETFILTER|(IP_NF_(IPTABLE|FILTER|TARGET_ULOG)))'

If they are not set as modules or compiled into the kernel you will need to change them and recompile the kernel. In menuconfig the following options need to be set:

Networking options > Network packet filtering
Networking options > Netfilter Configuration > IP tables support
Networking options > Netfilter Configuration > Packet filtering
Networking options > Netfilter Configuration > ULOG target support

You might also want to verify that iptables is compiled with ulog support.

2.1. Install MySQL

You can go directly to 2.2. if you have MySQL already setup. Otherwise:
apt install mysql
/etc/init.d/mysql restart
chkconfig mysql on

If you are using SuSE and do not have apt4rpm installed on your system, I highly recommend that you do so, as it will greatly simplify your package management issues.

You also need to set a password for the MySQL root user:

mysqladmin -u root password 'yourpassword'

2.2. Initialize the database

Type in:
mysql -p -u root
then enter your password at the prompt. Once you have logged into your MySQL database, enter the following commands to prepare the database to receive firewall logs from ulog.
create database ulogdb;
use ulogdb;
source /path/to/nulog/scripts/ulogd.mysqldump;
grant select,insert,update,drop,delete,create temporary tables, on ulogdb.* to ulog@localhost identified by 'ulogpass';
flush privileges;
quit;
So what happened here?

3.1. Install ulogd

You will need to install the logging daemon ulogd:
apt install ulogd-mysql

3.2. Configure ulogd.conf

Edit /etc/ulogd.conf to match what we set up previously:

[MYSQL]
table="ulog"
pass="ulogpass"
user="ulog"
db="ulogdb"
host="localhost"

You should change the password “ulogpass” to the password you set in the GRANT command in your MySQL database. Now uncomment the following line to send the data to MySQL:

plugin /usr/lib/ulogd/ulogd_MYSQL.so

and comment out the following two lines to prevent logging to a text file:

#syslogfile /var/log/ulogd.syslogmenu
#plugin /usr/lib/ulogd/ulogd_LOGEMU.so

Now restart the ulogd daemon and set it to be automatically started at boot time with chkconfig:

/etc/init.d/ulogd restart
chkconfig ulogd on

4. Redirect iptables Logging

The following sed command switches all your iptables rules to log through ULOG, we will assume that you store your iptables ruleset in a file called “iptables” (usually in /etc/sysconfig/ or /var/lib/)

sed 's/LOG/ULOG/'; /etc/sysconfig/iptables > /etc/sysconfig/uiptables
iptables-restore < /etc/sysconfig/uiptables

You are now all set up! All the logs from your firewall are now being logged in your MySQL database. Don't forget to update your firewall startup script so the new iptables are taken into account.

5. Import Your Old Logs

So far, so good, but you probably would like to have your old logs in MySQL also. Here is a little perl script to allow you to import your old text logs to MySQL. Some of the regexps are reused from adcfw-log. You can usually find your netfilter logs in /var/log/firewall-XXXXXX.gz or /var/log/messages-XXXXXX.gz. To import:

gunzip -c /var/log/firewall-XXXXXX.gz | nf2sql.pl
Repeat for each of your other log files. To process a current log file (or other uncompressed log file) such as /var/log/messages or /var/log/firewall:
nf2sql.pl < /var/log/messages
That's it!

6. Analyze the Results

To analyze your logs in MySQL you can use nulog or webfwlog

Links

This article was partly inspired by this article (only available in Spanish).

The original ulog page can be found here.

If you want to push it further and log all system messages to MySQL, you can take a look at this HOWTO setup PHP syslog-ng wiki entry.

Here is a reason to move away from the usual text file logging.

If you do not have your iptables set already, you can easily build a good ruleset with shorewall, firehol or firestarter.


[BIO] A. N. Onymous has been writing for LG since the early days - generally by sneaking in at night and leaving a variety of articles on the Editor's desk. A man (woman?) of mystery, claiming no credit and hiding in darkness... probably something to do with large amounts of treasure in an ancient Mayan temple and a beautiful dark-eyed woman with a snake tattoo winding down from her left hip. Or maybe he just treasures his privacy. In any case, we're grateful for his contribution.
-- Editor, Linux Gazette

Copyright © 2005, Anonymous. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

Using the GNU Compiler Collection (Part2)

By Vinayak Hegde

In the last article we saw some of the basic GCC options, and noted that it supports several CPU architectures. One of the topics we will cover in this article is how to turn on optimizations for different architectures and what happens when they are turned on. We will also look at some other nifty tricks which we can do with GCC.

Adding symbols for profiling and debugging

Profiling is a method of identifying sections of code that consume large portions of execution time. Profiling basically works by inserting monitoring code at specific points in the program. This code can be inserted by using the -pg option of GCC.

When debugging we need extra information added to the binaries. Programs compiled with the -g flag additional information which can be used by gdb (or other debuggers) is added to the binary. This increases the size of the binaries but is necessary for debugging. When compiling debugging binaries we should turn off all optimization flags. GCC can add debugging information in several different formats such as stabs, dwarf-2 or coff format.

$ gcc -g -o helloworld helloworld.c #for adding debugging information
$ gcc -pg -o helloworld helloworld.c #for profiling

Programs compiled with profiling and/or debugging turned on are usually referred to as debug binaries, as opposed to production binaries which are compiled with optimization flags.

Monitoring compilation times

We previously saw that the compilation of the program to get an executable binary consists of different phases. Each of the main compile stages (compiling to assembly language, assembling and linking) is done by different executables (e.g. cc1, as and collect2). We use the -time option to GCC to get a breakdown of the time required for each stage.

$ gcc -time helloworld.c
# cc1 0.02 0.00
# as 0.00 0.00
# collect2 0.04 0.01

We can also gather more fine-grained statistics about the various stages of the compiler cc1 using the -Q option. This shows the real time spent as well as time spent in userspace and kernel modes.

$ gcc -Q helloworld.c
 main
Execution times (seconds)
 preprocessing         :   0.00 ( 0%) usr   0.00 ( 0%) sys   0.24 (38%) wall
 parser                :   0.01 (50%) usr   0.00 ( 0%) sys   0.02 ( 3%) wall
 expand                :   0.00 ( 0%) usr   0.00 ( 0%) sys   0.03 ( 5%) wall
 global alloc          :   0.00 ( 0%) usr   0.00 ( 0%) sys   0.03 ( 5%) wall
 shorten branches      :   0.00 ( 0%) usr   0.00 ( 0%) sys   0.04 ( 6%) wall
 symout                :   0.00 ( 0%) usr   0.00 ( 0%) sys   0.01 ( 2%) wall
 rest of compilation   :   0.01 (50%) usr   0.00 ( 0%) sys   0.00 ( 0%) wall
 TOTAL                 :   0.02             0.00             0.64

GCC Optimizations

Before we move on to optimizations, we need to look at how a compiler is able to generate code for different platforms and different languages. The process of compilation has three components:

There are two kinds of optimizations possible - optimizations for speed and optimizations for space. In an ideal world, both would be possible at the same time, (Actually some optimizations do both - such as common sub-expression elimination). More often than not optimizing for speed increases the memory footprint (the size of the program loaded in memory) and vice versa. Expanding functions inline is a good example of this case. Inlining functions reduces the overhead of a function call but ends up replicating code wherever the inline function has been called, thus increasing the size of the executable. Turning on optimizations will increase the compilation time as the compiler has to analyze the code more.

GCC offers four optimization levels. These are specified by the -O<Optimization Level> flag. The default is no optimization or -O0 (notice the capital O). Various optimizations are turned on by the each of the different levels (-O1, -O2 and -O3). Even if we give higher optimization levels such as -O25, they have the net effect of enabling the highest level of optimizations (-O3). In addition to these four optimization level there is another optimization level -Os which enables all the optimizations for space as well as those optimizations which do not increase the size of the code but give speed improvements. In -O1 optimization level, only those optimizations are done which reduce code size and execution time without increasing compilation times significantly. In -O2 optimization level, those optimizations which do have a space-execution time tradeoff are done. Almost all optimizations are turned on by -O3 optimization level but compilation time might increase significantly by turning it on.

$ gcc -O3 -o hello3 helloworld.c
$ gcc -O0 -o hello0 helloworld.c

$ ls -l
-rwxr-xr-x  1 vinayak users    8722 2005-03-24 17:59 hello3
-rwxr-xr-x  1 vinayak users    8738 2005-03-24 17:59 hello0

$ time ./hello3 > /dev/null
real    0m0.002s
user    0m0.001s
sys     0m0.000s

$ time ./hello0 > /dev/null
real    0m0.002s
user    0m0.000s
sys     0m0.003s

As seen above, compiling the program with -O3 optimization level reduces the size of the executable compared to the compiling with -O0 optimization level (no optimization)

It is also possible to have CPU or architecture specific optimizations. For example a particular architecture may have numerous registers. These can be utilized by the register allocation algorithm intelligently so as to store temporary variable between calculation to minimize cache and memory accesses thus ensuring considerable speed ups in CPU-intensive operations. Some of the platform specific optimization flags can be done using -march=<architecture type> or -mcpu=<CPU name>. For the x86 and x86-64 family of processors, -march implicitly implies -mcpu. Some of the architecture types options in this family are ix86 (i386, i486, i586, i686), Pentiumx (pentium, pentium-mmx, pentiumpro, pentium2, pentium3 ,pentium4) and athlon (athlon, athlon-tbird, athlon-xp, opteron). But executables built with platform specific flags may not run on other CPUs. For example, executables generated with the -march=i386 will run on i686 platform because of backward compliance of the platforms. However, executables generated with the -march=i686 may not run on the older platforms as some of the instructions (or extended instruction sets) do not exist on older CPUs. If you use the Gentoo Linux distribution, you might be already familiar with some of these flags.

$ gcc -o matrixMult -O3 -march=pentium4 MatrixMultiplication.c #optimise for Pentium 4
$ gcc -o matrixMult -O3 -march=athlon-xp MatrixMultiplication.c #optimise for Athlon-XP

Also you can give specific flags for optimizations at the command line such as -finline-functions (for integrating simple functions into their callers) and -floop-optimize (to optimize loop control structures). The important thing to remember is that the order of the flags on the command line matters. The option on the right will override ones on the left. This is a good way to choose particular options without cluttering the compilation command line. It is also possible to do it for platform-specific optimizations.

$ gcc -o inlineDemo -O3 -fno-inline-functions InlineDemo.c
$ gcc -o matrixMult -march=pentium4 -nosse2 MatrixMultiplication.c

In the example, the optimization level -O3 will enable inlining of functions - the effect is same as -O2 -finline-functions -frename-registers. So if you have the option -fno-inline-functions on the right on the command line, it will disable inlining of functions. This command will turn on all the Pentium4 specific optimizations but code generated will not contain any MMX, SSE and SSE2 instructions.

Supported options

All the options supported by GCC on your machine can be seen by giving the following command:

$ gcc -v --help | less

This will list out all the different options that are supported by GCC and the processes invoked by GCC on your machine. This is a pretty huge list and should contains most of the options discussed by us in this article and the earlier one in this series and more.

Conclusion

In this article, we looked mainly at the various GCC optimizations options and how they work. In the next part of this series we will look at another development tool called make used for building big projects.

Resources


[BIO] Vinayak Hegde is currently working for Akamai Technologies Inc. He first stumbled upon Linux in 1997 and has never looked back since. He is interested in large-scale computer networks, distributed computing systems and programming languages. In his non-existent free time he likes trekking, listening to music and reading books. He also maintains an intermittently updated blog.

Copyright © 2005, Vinayak Hegde. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

A New Scanner with XSANE and Kooka

By Edgar Howell

Introduction

Long before my old copier (a multifunction unit) died, I was looking for a long-term, non-combo solution. For quite a while I have wanted a flatbed scanner to enable copying not only something from the revenuers for our accountant but also the occasional page from a book or magazine without having to destroy the publication first, so I really wanted a single-function device.

I also did not want another something requiring ink or toner or whatever. Besides, a satisfactory network printer was available. All that was really needed was a source of input, i.e. a scanner.

After much researching of product reviews and a bit of ping-pong with http://SANE-project.org, I settled on something with which I am quite happy.

The Scanner

For what it is worth, my choice was a Canon LiDe 20 which cost €55 (about US$70 last summer) from Amazon - no shipping charges, delivery in about 10 days. The LiDe 30 is supposed to produce slightly better quality than the LiDe 20, but not in proportion to the difference in price.

This is not heavy-duty hardware, and it is not intended to be out for lots of use every day. Instead it has a foot that lets it be stored upright out of the way on the floor next to the PC when not in use. It connects to the PC through USB connection when in use. It seems to be ideal at home or in a small office.

But the hardware isn't as important as the software and the process of obtaining and using it. And thanks to the help of the folks at SANE-project.org, selecting hardware was almost a trivial exercise and using it equally easy. They have an incredible amount of information on the performance of various scanners under GNU/Linux. I did glance at the literature that came with the hardware but none of the information or software provided by the manufacturer was needed.

Software Choices

In place of the manufacturer's software, I installed XSANE (the graphical front-end for SANE). Since SuSE had Kooka (the KDE equivalent of XSANE), I installed that as well. Because much of my research had been at the SANE project, I planned on using XSANE, but then decided "why not Kooka?"

Usually GUI-based programs are flexible, have lots of choices and can be tweaked on-the-fly as needed. They don't necessarily work well with repetitive, run-of-the-mill situations, however. What I really wanted was exactly that: a quick copy off to the network printer. So, OK, it isn't yet apparent how these guys can help, but let's check them out.

As it turns out, for me at least, Kooka has one advantage: it can print without having to first save a file, then start some other program to do the actual printing. That makes it simple to use it with the scanner and network printer as an alternative to a single-purpose copier.

Running the Programs

When started, both ask for confirmation of the device to use. But Kooka starts a single window, subdivided into 3 sections for a list of recent scans, settings, and preview/current scan, while XSANE starts 4 separate windows (there are more available) providing preview and the like. Since I am a keyboard aficionado from the days of the typewriter and not a mouse-user at all, I prefer Kooka's behavior - XSANE turns Alt-TAB (tabbing among windows on the current desktop) into a bit of a farce.

Both allow selecting either color, black-and-white, or grayscale. They also have buttons for the usual manipulations to rotate or mirror and the like. And they each show you the results of scanning as well as offering a preview scan, again Kooka within a section inside the window, XSANE starting another window for this.

To me it seems pretty clear that printing was never really high priority in either software package, at least not in the sense of using a scanner as part of a replacement for a copier. Kooka has a print button, and XSANE allows access to Gimp with which one can print (at least in theory; this is something I haven't yet pursued).

Kooka also turns about 2/3 of a printed page into almost 14MB for the postscript printer. Patience... The quality isn't bad, though. To be fair, the quality on the CRT is far better than on the printer, not a photo-quality device.

Since I am very new to the scanning game, I don't yet appreciate or even understand much of the available functionality, but there is lots more!

Conclusion

When all is said and done, I am extremely impressed with what SANE, the two GUI interfaces, and the hardware accomplish. Super quality on a CRT, ideal for sending to someone as an attachment to an e-mail -- although I do need to play around with formats to find something less than 1MB for my analog modem.

Recently a scan of less than one printed page from a newspaper produced a JPEG of less than 150K, quite acceptable for forwarding as an attachment. This wasn't even on my original wish-list!

As far as the occasional "quick copy" is concerned there is a trade-off. It requires some time to move the scanner to a free area somewhere and attach it to a USB port. And then to start one of the GUIs. But copying has never been a frequent requirement in this SOHO and I have regained a bit of desk space.

This is clearly not the way to go for everybody, especially not if you are short on space, such as in student housing or a small apartment. It is ideal for a small office without continual need for copying. In any case, do check out SANE before you look for scanners or copiers.

Disclaimer

Use of scanners or copiers involves risk of violation of copyright. To my knowledge in some jurisdictions there is the concept of "fair use" which includes quotes of ill-defined length. When I copy something it is either (1) for my own personal temporary use because I don't want to risk loss of or damage to the original or (2) to forward to a friend along with attribution, which I consider not only "fair use" but advertisement for the source. I am not a lawyer, so use your own judgment.


[BIO] Edgar is a consultant in the Cologne/Bonn area in Germany. His day job involves helping a customer with payroll, maintaining ancient IBM Assembler programs, some occasional COBOL, and otherwise using QMF, PL/1 and DB/2 under MVS.

(Note: mail that does not contain "linuxgazette" in the subject will be rejected.)

Copyright © 2005, Edgar Howell. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

The Basics of DNS

By Rick Moen

This article is a result of a discussion on the Answer Gang list, in which Rick had brought up some interesting and common problems with DNS (Domain Name Service). Since DNS is 1) a critical part of the Internet infrastructure, 2) one of the most important - and yet very easy - services Linux users can provide for each other, and 3) seemingly poorly understood and seen as Deep Wizardry by most Linux enthusiasts, I asked Rick to expand on the issue. His response follows.
-- Ben

Quoting Benjamin A. Okopnik (ben@linuxgazette.net):

> Rick, is there a _simple_ way to do basic DNS service, or is it an "all
> or nothing" sort of deal?

As a matter of logical categories, I can spot four distinct categories of "DNS service": three of the four are dead simple. The fourth isn't too difficult if you can refer to example zonefiles as your model. Let's run through them, in turn, from simplest to most complex.

1. Recursive-resolver nameserver

The idea here is that you want to run a local nameserver for its caching abilities, but you're not serving up authoritative DNS information of your own or for someone else. You just want local machines to have somewhere _local_ to query for name lookups, rather than having all queries go across the wire to your ISP or elsewhere -- in order that, most of the time, the answer's already in your nameserver's cache because some other local machine also made the same query in the recent past.

How do you enable it, you ask? You just turn on your nameserver. Conventional DNS daemons (BIND9, MaraDNS, Posadis, PowerDNS, Twisted Names, Yaku-NS) default to that sort of service, so you just switch them on, and they work.

It's that simple.

Oh, and on the client side, you'll want to configure your various machine to consult that nameserver in the future, via "nameserver" entries in their /etc/resolv.conf files (the configuration file for a client machine's "resolver library", the DNS client that on typical Linux machines is built into 'glibc'). For client machines that are on DHCP, you can automate this configuration via a suitable entry in dhcpd.conf.

2. Caching forwarder nameserver

This type of service is only subtly different: Basically, the nameserver daemon is one that lacks the smarts to, by itself, recurse through DNS namespace on queries. Instead, it forwards all queries it receives to a full-service nameserver elsewhere, which does the job. Your local (caching forwarder) nameserver merely caches recently received answers in case they're needed again, and of course ages the cache. On the plus side, avoiding the very difficult coding problems posed by _not_ handling recursive-resolver functionality means the daemons can be very small and secure. Examples include dproxy, Dnsmasq, DNRD, ldapdns, and pdnsd. pdnsd is particularly popular for really small networks and laptops, in particular because it stores its information in a disk-based cache that is (therefore) non-volatile.

How do you enable it? You put the IPs of one or more "upstream" full-service nameservers in its configuration file (to tell it where to forward to). Then, you turn it on, and it does its thing without further fuss.

Again, it's that simple.

3. Secondary authoritative nameserver

This is the case where your friend Alice Damned <alice@mydamnedserver.com> asks you "Will you help me by doing secondary nameservice for mydamneddomain.com?" You respond with, "Yes. My nameserver, ns1.example.com, is at IP address 10.0.42.1. Please add that to your allowed-transfer list, add an appropriate NS line to your zonefile, and make my IP authoritative -- and we'll be in business." (Your telling Alice that is kind of superfluous, actually, in the sense that those things are her problem to figure out and implement, but let's suppose you're trying to be helpful.) She also should have been equally helpful by telling you what IP address her primary nameserver lives on. If not, you do this, to find out:

$ dig -t soa mydamneddomain.com +short

The global DNS should return with a hosthame plus other details (that you can disregard, for this purpose) from Alice's domain's Start of Authority (SOA) record, something like:

ns1.mydamneddomain.com. alice.somewhere-else.com. 2005112200 7200 3600 2419200 86400

Which tells you that the primary DNS is claimed to be provided by ns1.mydamneddomain.com. Use the 'host' command to get the corresponding IP address. Let's say 'host' returns IP address 10.0.17.1 for that hostname.

How do you enable it? If you already are running a nameserver capable of authoritative service (let's say, BIND9), then you need to lavish five minutes of your time on a new "stanza" (paragraph) in your nameserver's main configuration file, instructing it to (also) do secondary nameservice for this domain. Again, using BIND9 as an example, one would add this to '/etc/bind/named.conf' (or wherever you put local additions, e.g., '/etc/bind/named.conf.local'):

//For Alice Damned <alice@somewhere-else.com> 212-123-4567 cellular
zone "mydamneddomain.com" {
        type slave;
        allow-query { any; };
        file "/var/cache/bind/mydamneddomain.com.zone";
        masters {
        10.0.17.1;
        };
};

Notice the comment line: You want to have on hand reliable means of contacting Alice in case you need to talk to her about problems with her domain -- and ideally means of communication that do not go through the domain in question (as "Your domain is offline" mails don't work too well when they're blocked by the fact that the domain's offline).

In the case of BIND9, you can make your nameserver reload (or load) a single zone such as mydamneddomain.com using BIND9's 'rndc' (remote name daemon control) administrative utility, as the root user:

# rndc reload mydamneddomain.com

You should, if everything's configured right, now see your local cached copy of Alice's primary server's zonefile (her domain's DNS information) pop into (per the above) directory /var/cache/bind/, as file mydamneddomain.com.zone. The moment you see that, you're done: The contents and configuration of the zonefile are strictly Alice's problem.

If you don't see a copy of the zonefile appear (that copy operation between nameservers being referred to as a "zone transfer"), then either you've made some silly error, or Alice's nameserver isn't willing to send yours the zonefile because she made some silly error. One of you will probably find a clue in his or her '/var/log/{daemon.log|messages}' file, fix the silly error, reload the zone or restart the nameserver as required, apologise, and move on.

The nice thing about setting up secondary DNS is (1) it's pretty much a set-up-and-forget affair on your end, and (2) it's the other person's (Alice's) job to notice most sorts of problems. Moreover, it's usually be her screw-up. So, doing secondary is an easy way to help a friend, and involves only a tiny amount of one-time work.

4. Primary (master) authoritative nameservice.

This is the exception, the case where you actually need to know what you're doing: This is where you're Alice. You have to maintain the zonefile, which then gets propagated to all your secondaries via zone-transfer mechanisms. You have to check on your secondaries from time to time, making sure they haven't shot you in the foot by, e.g., forgetting to carry forward that "slave" stanza when they rebuild their servers.

How do you enable it? Here is Alice's BIND9 "stanza" that operates her end of the mydamneddomain.com operation:

//For myself
zone "mydamneddomain.com" {
        type master;
        allow-query { any; };
        file "/etc/bind/mydamneddomain.com.zone";
        allow-transfer {
        //Joe Gazettereader <joe@example.com>, 212-765-4321 cellular
        //ns1.example.com, is:
        10.0.42.1;
        };
};

Again, notice the comment lines, basically notes that Alice keeps for her reference in case she wants to reach Joe in a hurry about him shooting her domain in the foot. The "allow-transfer" list is the list of IPs that are permitted to transfer (pull down) Alice's zonefile, just as the "masters" list in Joe's earlier stanza listed the IPs of machines that Joe's secondary service expects to be able to pull down transfers from.

That leaves only the other difficult problem, which is the composition and maintenance of Alice's zonefile. I'm going to wimp out and claim it's out of scope for a brief article on simple DNS service, but will point out that I publish a set of example BIND9 zonefiles and configuration files that many people have used as examples to work from: http://linuxmafia.com/pub/linux/network/bind9-examples-linuxmafia.tar.gz

Did I say "the other difficult problem"? Oops, there are more: as publisher of primary (master) authoritative nameservice, you need to be concerned not only that your domain's zonefile contents are correct, but also that your domain itself is set up correctly, at your domain registrar -- including enumerating, there, all of your domain's nameservers to make them "authoritative" (i.e., tagged as a reliable source of information on the domain's contents, as opposed to just caching other nameservers' information if/when it happens to pass through). Getting your domain records wrong can have a variety of ill effects, and I can only recommend that (as with the finer points of zonefile contents) you ask a knowledgeable person, maybe in your user group, to check your work. Along those same lines, by all means use R. Scott Perry's excellent DNS-checking CGI scripts at http://www.dnsreport.com/, to check (in a single, amazingly useful report) both your domain records and your (in-service) zonefiles.

It's important to note that there are many good nameserver daemons, other than BIND9 -- which is important for historical reasons, but has the annoying problems of being, as I say in my list of all known DNS server programs for Linux, "a slow, RAM-grabbing, overfeatured, monolithic daemon binary". That list is in my knowledgebase at "DNS Servers" on http://linuxmafia.com/kb/Network_Other/, and contains a number of good choices for particular DNS situations. My page aspires, among other things, to identify which type of the four classes of service each package can do. I hope it's useful to people.

A wide variety of special-purpose primary-nameservice configurations are possible, such as running a deliberately non-authoritative nameserver (not shown in either your domain records or your zonefile) to provide master DNS service from a protected machine inaccessible and unadvertised to the public -- but details are beyond this brief article's scope.


[BIO] Rick has run freely-redistributable Unixen since 1992, having been roped in by first 386BSD, then Linux. Having found that either one sucked less, he blew away his last non-Unix box (OS/2 Warp) in 1996. He specialises in clue acquisition and delivery (documentation & training), system administration, security, WAN/LAN design and administration, and support. He helped plan the LINC Expo (which evolved into the first LinuxWorld Conference and Expo, in San Jose), Windows Refund Day, and several other rabble-rousing Linux community events in the San Francisco Bay Area. He's written and edited for IDG/LinuxWorld, SSC, and the USENIX Association; and spoken at LinuxWorld Conference and Expo and numerous user groups.

His first computer was his dad's slide rule, followed by visitor access to a card-walloping IBM mainframe at Stanford (1969). A glutton for punishment, he then moved on (during high school, 1970s) to early HP timeshared systems, People's Computer Company's PDP8s, and various of those they'll-never-fly-Orville microcomputers at the storied Homebrew Computer Club -- then more Big Blue computing horrors at college alleviated by bits of primeval BSD during UC Berkeley summer sessions, and so on. He's thus better qualified than most, to know just how much better off we are now.

When not playing Silicon Valley dot-com roulette, he enjoys long-distance bicycling, helping run science fiction conventions, and concentrating on becoming an uncarved block.

Copyright © 2005, Rick Moen. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

DNS definitions

By Mike Orr (Sluggo)

DNS administrators often speak of master/slave servers, primary/secondary servers, and authoritative/non-authoritative servers. These do not all mean the same thing but are often confused, both due to ignorance and because the official usage has changed over time. So the person you're speaking with may match any term with any of the meanings below, and you'll have to figure out from context what he means. This also means you should explain the term with anybody you're speaking with, or at least put a few words of context so they know which meaning you intend. Note that all these terms are domain-specific. A server can be master for one domain while simultaneously being slave for another domain.

master/slave

A master server knows about a domain from its own configuration files. A slave server knows because a master has told it. The slave is configured to retrieve that particular domain from a certain master, either through a DNS zone transfer or out-of-band (via 'rsync' or another mechanism.) Master/slave is a private relationship between the servers; neither the registrar nor the public know which IP is in the slave's configuration file, or even that it is a slave. A slave's "master" may in fact be slave to another master.

authoritative/non-authoritative

An authoritative server is listed at the registrar as having the official information for that domain. A non-authoritative server has the information because it earlier asked an authoritative server and cached the answer. You might say, "All slave servers are non-authoritative," but this is misleading. Slave servers contact their masters directly, while non-authoritative servers query the DNS hierarchy.

primary/secondary

These unfortunate terms were used for master/slave in earlier versions of BIND. However, some people think the primary is the first nameserver IP listed at the registrar, and any others others are secondary. In fact, all the nameserver IPs are equal and "authoritative"; the first one does not have a special status. Still other people think primary means the nameserver listed in the zonefile's SOA header, and others think primary means "the domain I personally edit". So avoid the terms primary/secondary. If you do use them (and it's hard not to let them slip out), take care to explain what you mean.

When I originally set up a domain for a nonprofit organization, I thought the first IP listed at the registrar had to be a master, and the others had to be slaves or the zone transfers wouldn't work properly. This turned out to be hogwash. A "hidden master" is actually quite common. That's where the real records are kept at a private or unadvertised server, and all the authoritative servers are slaves. This protects you from attacks: the cracker can get the money but he can't get the family silverware.

A question that comes up in those cases is "what value do I put in the SOA record?" (the item at the top of a DNS zone that tells which computers have the original configuration data). Traditional practice is to list the masters, but that is what you would not do if you really wanted to hide the masters. No DNS program actually uses the SOA value for anything as far as we know; it's more a note to humans than anything else, so you can use it to cue yourself, or your fellow system administrators, in whatever way you prefer.


picture Mike is a Contributing Editor at Linux Gazette. He has been a Linux enthusiast since 1991, a Debian user since 1995, and now Gentoo. His favorite tool for programming is Python. Non-computer interests include martial arts, wrestling, ska and oi! and ambient music, and the international language Esperanto. He's been known to listen to Dvorak, Schubert, Mendelssohn, and Khachaturian too.

Copyright © 2005, Mike Orr (Sluggo). Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

IT's Enough To Make You Go Crazy

By Pete Savage

I was warned by several people not to write this article whilst angry, for fear of it turning into a scroll of unspeakable darkness, conjured by a crazed homo sapiens wielding a typewriter. Dark materials such as these tend not to get published in a fine monthly such as the Linux Gazette. I decided from the outset that I would not to mention any names in this article. Not only to protect the identity of the incompetent and the blazingly stupid, but also to avoid the multitudes of replies standing up for company x, y or z. I suppose in some ways you could ask the question, why are you writing this article then? The answer? It just feels right, something I have to do. For too long I've been idly wittering on to people I meet about my love of open source software and the Linux OS, much to their disgust usually. Now is the time to stand up and be counted. Below is a summary of a particularly bad week, supporting common proprietary software.

And so the crazed homo-sapien began to type...

Earlier in the week I turned up at work and noticed a user's machine was running particularly slow. At first I considered the user's choice of OS as the limiting factor, but it was more than that. We're not talking "let's make a cup of coffee and go gossip by the photocopier" slow, more the kind of slow that causes a user to grow a beard - independent of their gender. I sat down and clicked my little rodent friend, only to be greeted by a barrage of web browser windows, offering me the latest in pharmaceutical developments at a fraction of the retail price. My credit card stayed firmly in its protective sheath and I resisted the temptation to shave more weight off my already skinny frame. You can already guess the issue that was at hand... SPYWARE and tonnes of it. The machine seemed to be sporting its own division of MI5, MI6, and CTU, along with a few KGB agents thrown in for good measure. I turned to the world wide web to download a well known anti-spyware product, clicked the download button, and wham, I was stopped quicker than a nitro equipped Ferrari F50.

The page you have requested is not accessible at this time.
The content group "Marijuana" has been filtered.

"For crying out loud," I screamed. I could tell this was going to be the beginning of a very long week. Our Internet systems are externally controlled and whilst I understand the reason for a good web filtering system, I do not believe that the company involved had started trading in illicit products. My heart sank.

I thought the next day would yield better luck. Perhaps I shouldn't have been so hasty. I had set a backup going using the OS's supplied archiving software at the end of last week. Having had little time on Monday to go and check its completion, I found a few spare minutes early on Tuesday whilst my brain was still actually firing neurons. I checked the report... data transferred: 3,995,363,231 Bytes. Seemed reasonable enough, not like I was going to sit down and count them. Elapsed time... 126 Days, 14 Hours, 6 Minutes, 23 Seconds. Excuse me?

Unless certain companies have now started incorporating space-time continuum shifting algorithms into their backup software, there was something seriously wrong. I mean, I'm not even sure how this would even work - although I have a few working theories.

  1. My software is trying to impress me.
    "Oh man I was so pooped Mr Network Manager, I squeezed 126 Days work into just 72 hours. Am I good or what? Go on, gimme that RAM upgrade you keep giving the Penguins."
    Verdict - I know this OS well. Not possible.
  2. We are backing up data close to light speed.
    Owing to the well know time dilation effect, maybe it's possible that the backup job was running in a different frame of reference to the rest of the machine.
    Verdict - Not feasible.
  3. We travel back in time and archive the data before it's even been created.
    "Yes that's right ladies and gentlemen, the new Backup 5000 will archive your work before you've even done it."
    Verdict - Begs the question, do I even need to bother doing the work in the first place. Can't I just restore it from the backup on the day it's due?
  4. There is a bug.
    Someone has screwed up what should be a relatively simple task. Something so simple, it was achieved, according to the Guinness Book of Records, over 500 hundred years ago: A working clock.

I learnt to tell the time at a fairly early age. Not as early as some of those super boff infants, who can point to the big and little hand at the age of 3, but simple time concepts, for example elapsed time, weren't exactly rocket science. It begs the question: if some programmers can't even tell the time, can we really entrust them with the safety and security of our collection of nuclear missiles? It does, however, explain the Y2K problem quite nicely. I can see the conversation now:

Person A: "So what do you do for a living?"
Person B: "I'm a programmer, I make computers do things."
Person A: "So you must be good at maths then?"
Person B: "If there's a sum I can't do, I haven't met it yet."
Person A: "What's 1999 + 1?"
Person B: "There are numbers after 1999? Bugger!"

By contrast, my Linux server seems to be able to tell the time quite well. Perhaps it's the collaboration of hundreds of Open Source programmers, who all, like me, were taught that elapsed time needn't be a complicated concept involving time machines. In fact, my backup routine doesn't even inform me of the time taken to perform the process. It doesn't need to. I don't have to acknowledge that it's done its nightly job every morning with the click of an OK button. I stick the tape in, punch a few keys, whisper the words "Go Crazy", and off she goes. That's probably the main difference between the two. I trust my linux server to do its job. I trust the others to need assistance.

Wednesday came and I began to lose all faith and sanity. This one's a tricky one to explain... suffice it to say we have a database storing information and a framework to access it. This was all purchased from a company, along with client access licenses (another great money making idea for the corporate fat cats) at some huge cost. My bosses decided to purchase another module for this framework. What happened next made me angry beyond belief:

I began to Moo [1]. Talk about milking it. Was it just me or did no one else see the gleaming pound/dollar/yen/other (delete whichever is appropriate) sign in the suppliers eyes?

The homo-sapien pauses for a breath and some malt-loafy goodness.

I'm not completely naive. I know some things must be charged for - Open Source food, anyone? I just feel that £7,000 and a further £1,500 a year for support for a module, that's right, folks - a module! - is about as sensible as drinking a vat of hydrochloric acid. One leaves a hole in your pocket, the other leaves a hole in your stomach. Go figure. Take into account also that this is an educational establishment, and you have a recipe for what I would consider severe injustice. Perhaps some of these companies are starting to claim back their future programmers wages already. Couple that with the fact that a developers license costs a mere £20,000 and my brain was just about ready to spread its wings and leave.

I mused for a while about whether there was an Open Source alternative. Google confirmed my suspicions. The problem being, from my experience, people just don't trust open source. According to one particularly uninformed individual I once met, it's evil. I begged to differ, but shortly afterward wished I hadn't. People seem to be scared of Open Source. The fact that the code has been checked by hundreds, if not thousands of programmers, and is available for all to see, is apparently a bad thing. I fail to see why. True, it's not without its problems, but the availability of free, customisable code wins over extortionate, closed source binaries any day. My advice: if you haven't already, try it.

24 hours later and I had decided not to keep track of time again until the weekend. I sat down to debug a particularly nasty CPU hog present on a user's laptop. After trying to ascertain the problem for what seemed like a few millennia, a strange thing happened. I was on my knees. That's right, I was actually begging my machine to show me what was happening. I'd given it the three-fingered salute, and it had thrown back something equally abusive, but I found myself pleading with it to give me some indication of what it was actually doing. The normal RAM bribes did nothing, and I was fresh out of ideas.

I can understand that for a desktop system, usability and nice, pretty, fluffy GUI's are almost mandatory, but there should, somewhere at least, be a method of viewing what's actually going on inside. My mind cut to a day dream. I imagined two monkeys. The first was sitting in front of a monitor with his glasses on, intently reading white text on a black screen whizzing by. He occasionally tapped a key or two and made satisfied "oooh ooh" noises at what he saw. Did I mention he was wearing a Debian tee-shirt and was called Henry? The second monkey sat on top of his monitor. The screen was showing a signal test, the case was off the side of the computer and monkey number two - I decided to call him Monty - was yanking various cards, chips and drives out of his machine, inspecting each one and giving it a gentle lick before throwing them at Henry. Cut to the end of the daydream, just before the credits:

Monty never did solve his problem and was committed to an asylum for the technically insane.

Henry recompiled his kernel that afternoon and went on to run a successful clinic, caring for the technically insane.

At this point in time, I felt a lot like Monty. Tired, lonely, and insane. Would licking my machine help? I quickly shunned the idea and went to lunch, in search of tastier things.

Had I been working at my linux box, I could have gathered all the information I wanted. A quick "top" command and I would have been presented with a list of all processes running on the system, their priorities, CPU Usage, Memory Usage, execution time, and maybe even been asked the question, "Do you want fries with that?" The main point to take away from this experience is that Linux is helpful. I can go as shallow or as deep into a problem as I like. From "It's broken", the kind of response a user normally gives, to performing an "strace -p " command and actually viewing the execution calls as and when they are happening. Granted it may seem more complicated at first, but why be like Monty when you can be like Henry?

Friday. The last day of the week. Excuse me for stating the obvious but at this stage even the facts seemed to be going astray. Surely, today would be kinder to me.

It didn't start well. Whilst munching on my breakfast, I decided to try to pay my gas bill. Having had little trouble paying it on-line before, I sat down and loaded faithful Firefox. After remembering my stupidly long and arduous authentication credentials [2], I was presented with my balance. I clicked on "Pay" and a feeling of darkness swept over me. I had a premonition that something was about to go horribly wrong; a common feeling in IT. The page cleared itself as it usually does and I waited and waited and waited. I looked under the desk to see if the little gremlins inside the router were flashing their torches at me, they were. I squinted back at the screen searching for the little spinning "loading" logo in the top right corner [3]. To my shock and horror it wasn't spinning. I refreshed the page; Same result. The page had apparently finished loading. How useful, a blank form with which to pay my bill. Do I sound sarcastic? I emailed the company to complain about a distinct lack of functionality, which I must admit I found difficult to describe.

Please describe the nature of the problem: "Nothing (Nuh-fing): A blank screen where a payment form should be."

Upon arriving home I loaded my inbox. I'm not quite sure what I was expecting, but something useful surely.

Dear Sir blah blah
I'm sorry but we currently only support Browser A.  
I suggest you use this to pay your bills in future.
We are thinking of introducing cross compatibility but not at this stage.
 -- Company X

Well shut my mouth. No, seriously, before the abuse just falls out. 100 Million people use the same browser I do! I guess that puts us in the minority, fellow fire-foxians! Excuse the sarcasm. I was immediately aware that the wall, which had previously been vertical and inanimate had started to hurl itself over my head. It took a few seconds to register that it was in fact ME banging MY head against the proverbial wall. This must be some kind of new stupidity warning device. The whole cross-compatibility support issue really bugs me. Why does the rest of the world insist on their own proprietary formats, when Open Source developers have been sharing theirs for years? Many Open Source packages will even read proprietary formats and often write to them too. OpenOffice is a great example. Not only can I work with the .odt format; a nice small file type, but I can also load the more common .doc format, and write to it. Did I mention I can sing a little ditty while I do all this too?

Several paracetamol later, I went up to bed and slept. Oh, did I sleep. I'd earned it. In short, I guess by writing this article I'm hoping some curios non-Linuxian/non-Open Sourcian will read it and think... there's another way? Yes, that's right, kiddies - you don't have be like Monty the monkey, you can solve problems the easy way. The brick wall needn't be your friend. You don't need a large bank balance to make it in life. You can have your cake and eat it. Linux is #1. Oh sheesh, who am I kidding, one monkey can't change the world!


[1] Mooing in an office environment is not generally advised, you tend to get strange looks followed by strange nicknames like Buttercup and Daisy. However, when the person calling you these nicknames is built like a Challenger II tank, you just simply smile and accept your shiny new office nickname. Keeps them from breaking their computers, I guess. Bless.

[2] It always fascinates me the information that companies choose to use for jogging our memory.

Pet's name - Because obviously pets are immortal and never die.
Favorite Colour - Another no-brainer, ask 100 people what 'turquoise' is. A large sea mammal? Generally users will either pick, Red, Green, Yellow, Blue, Black or White. If you get a really intelligent end user, we might get something as adventurous as purple, or sky blue. Heck - while we're at it, why not just go crazy? Here are a list of my favorites:

Favorite brand of hair conditioner.
Favorite insult.
Weight of my spouse in grams.
Cups of coffee consumed from birth till 21/09/2003

Enough!!

[3] It's when you've had a week like this that your brain starts to devolve. Complicated computing terms such as "browser processing indicator" are replaced by "little spinning loading logo", "food" becomes your "life force" and your "computer" becomes your "time-wasting device."


[BIO]

Pete has been programming since the age of 10 on an old Atari 800 XE. Though he took an Acoustical Engineering degree from the world-renowned ISVR in Southampton UK, the call of programming brought him back and he has been working as a Web developer ever since. He uses both Linux and Windows platforms. He still lives in the UK, and is currently living happily with his wife.

Copyright © 2005, Pete Savage. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

Ecol

By Javier Malonda

The Ecol comic strip is written for escomposlinux.org (ECOL), the web site that supports es.comp.os.linux, the Spanish USENET newsgroup for Linux. The strips are drawn in Spanish and then translated to English by the author.

These images are scaled down to minimize horizontal scrolling.

[cartoon]

Click here to see the full-sized image.

[cartoon]

Click here to see the full-sized image.

All Ecol cartoons are at tira.escomposlinux.org (Spanish), comic.escomposlinux.org (English) and http://tira.puntbarra.com/ (Catalan). The Catalan version is translated by the people who run the site; only a few episodes are currently available.

These cartoons are copyright Javier Malonda. They may be copied, linked or distributed by any means. However, you may not distribute modifications. If you link to a cartoon, please notify Javier, who would appreciate hearing from you.


Copyright © 2005, Javier Malonda. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 121 of Linux Gazette, December 2005

Tux