One of the tools I use as part of my day to day monitoring is called Logwatch; this extremely useful tool can be compiled to run on almost any Linux system and I have myself had it running on Debian, Gentoo and Ubuntu to name but a few distributions which include it within their software repository.

Installing Logwatch.

Assuming you have already su as root, are logged into as root, have correct administrative privilages or are simply using sudo at the beggining of every command; do the following to install logwatch:

apt-get install logwatch

Depending on your distribution of linux, apt-get might be replaced with aptitude, dkpg, Yum, up2date, etc. Once this has been done, logwatch should now be setup and the daily cronjob setup, however by default logwatch wont send emails and will instead output its daily reports to the root users local mail account. If you need to change this do the following:

nano /usr/share/logwatch/default.conf/logwatch.conf

Then find and change the following information within logwatch.conf:

Output = mail Format = html MailTo = test@email.com

Once you have done that, save and exit the file (Ctrl + O for save and Ctrl + X for exit). Next you need to update the cronjob and tell it to email the daily report to your email address by doing the following: first open the daily cron00logwatch:

nano /etc/cron.daily/00logwatch

Then find the line begining with /usr/sbin/logwatch it should be below the line saying #execute. By default it will mailto root so you need to change it to your email address so it looks something like:

/usr/sbin/logwatch --mailto test@email.com

Once you have done all the above you will have logwatch installed and setup to email daily reports. The default reports include httpd (apache), users logging in through sshd, PAM errors and disk space alerts. Logwatch is incredibly configurable and you can find or write additional functionality to your requirements. Certain other packages such as fail2ban are also monitored by logwatch so you are able to monitor the number of failed and banned logins.

The quake2-gwt-port was released yesterday April the 1st and like many I was dubious as to its reality due to it almost being something too good to be true, however after three hours of pain-staking building I managed to finally get it working. Due to the difficulties I encountered getting it working both under Debian with the 2.6.26-2-686 kernel and those simple problems found when doing the same within OSX, I have documented my findings here to help others in their quest at deciphering googles (at the moment) poor documentation on getting this to work!

Disclaimer

The information contained within this article is a journalised version of the notes I made while I went through the process of installing quake2-gwt-port, they have been published here for informal purposes only and you follow them at your own risk. I can not be held liable if your following the below in any way breaks your computer, just because it didn’t break mine, doesn’t mean it wont break yours. In other words getting this to work on your system can only really be done if you know what you are doing.

Compiling quake2-gwt-port under Debian

The first problem I ran into when attempting this under debian on my test server was the fact that java was not previously installed so running the following command quickly sorted that out, however it did take a few hours of searching and downloading 100MB+ files from the official JAVA website before I noticed that the second post on the google code page gave the answer.

apt-get install openjdk-6-jdk

Once java has been installed you need to install vorbis-tools and lame as these packages contain dependences that enable audio within the port, however while vorbis-tools will install through apt-get with no problems, the same can unfortunately not be said for lame. To get apt-get to install lame I had to update my /etc/apt/sources.list file with an additional source as documented within the debian wiki here and then run the following commands to update apt-get cache and install lame and vorbis-tools.

apt-get update
apt-get install lame
apt-get install vorbis-tools

As a side-effect of the aforementioned additional source added to the apt-get sources list, apt-get kicks up the following fuss when you update it following the changes to /etc/apt/sources.list:

W: GPG error: http://www.debian-multimedia.org lenny Release: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY 07DC563D1F41B907
W: You may want to run apt-get update to correct these problems

I would ignore these, and simply remove the additional source from /etc/apt/sources.list once lame has been installed as we know it’s something we have done to cause the error, and its not stopped anything from working (yet, as far as I know).

Finally so that you can download the source from google code, you should also install mercurial, this can be done within debian and many other Linux distributions through the following apt-get command as documented here.

apt-get install mercurial

For some reason on my version of debian the shell variable JAVA_HOME was not set up by the installation of Java (openjdk-6-jdk), therefore I had to crack open nano and edit /etc/profile to include the following just before the line that reads unmask 022:

JAVA_HOME="/usr/lib/jvm/java-6-openjdk"
export JAVA_HOME

If JAVA_HOME is not set then the compile process of Quake II will fail and you would have waisted 12 minutes waiting for the command line to tell you that.

Once you have followed the aforementioned you should now have the following required packages installed mercurial, openjdk-6-jdk, lame and vorbis-tools. Your now ready to download and build Quake II for debian using the following simple steps:

1. Load up your shell and cd into your home directory or other such workspace where your comfortable and run the following command to get a fresh copy of the code from google code:
   hg clone https://quake2-gwt-port.googlecode.com/hg/ quake2-gwt-port
2. cd into the newly formed quake2-gwt-port directory
3. Run ./build-dedicated-server this will build the tools and the client and server code.
4. Next Run ./install-resources which will download, unpack, and convert the original Quake II demo resources, for reasons of copyright respect these have to be downloaded from their source and cant be packaged on google code.
5. Due to a glitch in the way the above two commands install the code, you then have to copy the created gwtquake from maven-build/server/target/gwtquake/war/gwtquake to /war/gwtquake using the following command: cp -r maven-build/server/target/gwtquake/war/gwtquake war/qwtquake
6. Everything is now ready to run using the command ./run-dedicated-server [port]

On my debian set-up I encountered several [WARNING] notices during the resource install but those where mostly related to character encodings and did not halt or impair the install. As I installed this on my server I have not been able to test it works within a browser due to a lack of time, however the install under osx is a much simpler affair.

Compiling quake2-gwt-port under OSX

Please Note: you must have OSX developer tools installed before you can compile anything.

As OSX does not come with vorbis-tools or lame installed you have to grab them yourself, fortunately with the fact that OSX is basically Linux with a pretty user interface doing so is very easy. To install the additional components required for the compiling of quake2-gwt-port under OSX I would recommend installing homebrew,over the much bloated MacPorts as given within the google code instructions. Installing homebrew is a very simple task indeed, first read through the online install instructions so you know what your doing then simply run the following two lines of commands:

sudo chown -R $USER /usr/local
curl -Lsf http://github.com/mxcl/homebrew/tarball/master | tar xvz -C/usr/local --strip 1

Once that has completed installing the required items for compiling the quake2-gwt-port is as simple as running the following lines of commands within OSX terminal:

brew install vorbis-tools
brew install lame
brew install mercurial

Once that is done you can continue with the instructions 1 to 6 above under the debian install guide.

Playing quake2-gwt-port under OSX

Hopefully the install went smoothly and you are now ready to play with your new toy, unfortunately just sending any old browser to http://localhost:8080/GwtQuake.html wont work as almost all modern day browsers are not compatible with HTML5, or if they are its only half an implementation due to the protocols still being on the drawing board. Two browsers that are compatible with the quake2-gwt-port are Google’s own Chromium browser and Webkit (click on links for download pages), which I presume to be the developers version of Safari due to its icon but I may be mistaken.

To get Webkit to play the game you first have to run the following command in terminal to enable webGL to be utilised by HTML5:

defaults write com.apple.Safari WebKitWebGLEnabled -bool YES

To get Chromium to play the game you have to first pass it a set of variables as you start it up, to do this, in your terminal cd to the applications folder where you have saved Chromium and run the following command to load Chromium with WebGL enabled:

./Chromium.app/Contents/MacOS/Chromium --enable-webgl --no-sandbox

Review of Game-play in Chromium and Webkit

At the moment I am unable to get the game to run in Chromium with sound, however with Webkit while audio played amazingly well as did single player, when I set it up to run the demo it crashed every time. The performance in Webkit was also noticeably better than in Chromium, with Webkit hitting 40fps at one point while Chromium lagged behind with an average of just 10fps! I have included below some screen shots of each browser playing the game and shall post updates when I get the version installed on my server working multi-player with some systems in my house.

If you have attempted getting this to work, like the article or have something to say, please add your comment below; I always try to respond to all comments posted here.

It seems that my Corsa has developed the same problem that many others have, which under the circumstances I can neither afford or live with. Its steering column has begun to prefer steering in one direction, while it is quite funny to think of a car having preferred directions giving it a more Herb-like feel, it is a little worrying when the steering locks up while steering around corners to the point I’m afraid it may not steer when I turn the wheel.

Although mechanically the locking up is simply the electric motor of the powered steering responding to incorrect torque readings which means it is unlikely that the cars steering will fail it is still something that’s on the back of my mind. Hopefully things will be fixed tomorrow when the mechanic comes round to have a look at it, I like knowing a mechanic whom does house calls – now if only we could have that kind of service with doctors again, like in those old days.

I got much of my infomation from talking to people in the know and the rest from reading this. Image Source: Giles.

I kept getting compile errors such as:

make -C src/ clean
make[1]: Entering directory `/share/r8101-1.012.00/src'
rm -rf *.o *.ko *~ core* .dep* .*.d .*.cmd *.mod.c *.a *.s .*.flags .tmp_versions Module.symvers Modules.symvers Module.markers *.order
make[1]: Leaving directory `/share/r8101-1.012.00/src'
make -C src/ modules
make[1]: Entering directory `/share/r8101-1.012.00/src'
make -C /lib/modules/2.6.26-2-686/build SUBDIRS=/share/r8101-1.012.00/src modules
make[2]: Entering directory `/lib/modules/2.6.26-2-686/build'
make[2]: *** No rule to make target `modules'.  Stop.
make[2]: Leaving directory `/lib/modules/2.6.26-2-686/build'
make[1]: *** [modules] Error 2
make[1]: Leaving directory `/share/r8101-1.012.00/src'
make: *** [modules] Error 2

These errors caused much head scratching in #debian before one member valdyn had a go at compiling it himself and sent me the driver as it would not compile on my machine. I am still not sure why it would not install however as I now have a compiled driver I don’t need to investigate any further into that. Now that I have the driver it needs to be moved into the following folder:

lib/modules/2.6.26-2-686/kernel/drivers/net

Then run the following commands to finalise the install:

depmod -a
modprobe -r r8169
modprobe r8101
lsmod | grep r81

The response to lsmod as shown below means that the driver was loaded successfully:

r8101                  26992  0

I double checked this by running ifconfig which gave the response below showing no dropped packets.

RX packets:6232 errors:0 dropped:0 overruns:0 frame:0
TX packets:5214 errors:0 dropped:0 overruns:0 carrier:0

While I am not sure if the following is required after having deleted the module I wanted to make sure that it would not reload on boot so I have added it to the modules blacklist via opening /etc/modprobe.d/blacklist and appending the following lines to the bottom:

# blacklist the default driver for realtec
blacklist r8169

Once all that is done I can now reboot the server and it load the correct driver for the realtec network chip. I have attached below a copy of the compiled driver for those whom cant compile it themselves.

Download Driver

As the bug has been fixed in the package already shipped with current versions of linux, there should be no reason to download this driver. If you do still require a compiled version please get in touch via comments below and I will re-upload the driver to here.

It looks like the fan is a 40x40x10mm standard build so I am going to be looking for something like the SUNON KDE1204PFV2.

I found infomation on this here, here, here and here.

In Soviet Russia, bulb lights you, well ok not the best Soviet Russia joke but it was worth the shot. The picture above is however no joke, as bad as it is, this is the first light bulb made available for sale in Soviet Russia. According to the images source they had Lenin and Stalin silouete cutouts as glowers within the glass bulb and where considered at the time to be a very classy present with the first batch being presented to the delegates of Soviet parliament of 1935. Now they are considered very valuble anteques and can sell for upwards of $2000USD in Moscow antique shop.

First things first monitorix requires rrdtool and librrds-perl, these can be installed with the following lines in shell:

apt-get install rrdtool
apt-get install rrdtool-perl

Once done and it should only take a few seconds, I moved o to the download page at the monitorix project. The latest package to date for Debian is the any platform linux distro monitorix-1.2.1.tar.gz. I moved to my user space and the used wget to download the package to the server without having to mess with ftp, then I extracted the archive using gzip:

cd /
wget http://www.monitorix.org/monitorix-1.2.1.tar.gz
gzip -dc ./monitorix-1.2.1.tar.gz | tar -x

Unfortunately there is no automatic apt-get script for Debian, however the monitorix group have been kind enough to simplify things by creating an install script which is as simple to run as:

cd monitorix-1.2.1
./install.sh

After that the install script asks you which system your installing it on press three for Debian (or if your reading this for another distro which ever number equals your os), next check the paths it gives and press enter if ok. I don’t know what they should be and no manual is written that will tell me so I press enter and hope for the best, you’re then issued with a Last chance to stop the installation cross your fingers hit ‘y’ and press return. If all goes well you should end up with something that looks like the following:

Installation successfully finished.

Now you have to edit the configuration file /etc/monitorix.conf
in order to setup correctly Monitorix to your operating system.

Then you'll can start Monitorix executing the init script:

/etc/init.d/monitorix.sh [start|stop|status]

and finally go to http://localhost/monitorix/ to start to see results.

NOTE: As a crond-based application, the root user will receive emails
about Monitorix execution errors. Please check the root email..

If so your now ready to setup monitorix, there are few things you need to set here unless your a more advanced user an in that case why are you even reading this, you guys are like god. In any case I use trusty old nano for the task of text editing and then start the service.

nano /etc/monitorix.conf
/etc/init.d/monitorix.sh start

Now if your server isn’t running SE-linux then when you go to http://your.domain/monitorix/ and see pretty things, however your like me and unlucky enough to be running a secure server the damn thing don’t work and simply throws up the following:

Starting Monitorix: ERROR: while creating /var/lib/monitorix/cpu.rrd: creating '/var/lib/monitorix/cpu.rrd': No such file or directory... Failed

Where as previous people seem to stop there and either disable SE-Linux to get it working, something which I believe is akin to running windows 3.1 on an unsecured network – you just shouldn’t do it, even if it works. I simply logged in as root and created the directory it so eagerly wanted to create using mkdir before retrying the start the process and being hit again with a different error:

mkdir /var/lib/monitorix/
 /etc/init.d/monitorix.sh start
Starting Monitorix: print() on closed filehandle OHTML at /usr/bin/monitorix.pl line 705.
readline() on closed filehandle IHTML at /usr/bin/monitorix.pl line 797 ... Ok

That Ok at the end doesn’t give me much hope, however it is a sign that something has worked and in a way which didn’t involve disabling server security. The problem now however is that Apache doesn’t have enough file rights to modify the $BASE_WWW directory, the $ meaning its a fixed variable not the directory name itself. I don’t know how to find this out so a further google search is needed, it would appear that the issue has been fixed for ubuntu users via here, so I believe I should be able to follow their instructions to fix this issue. The problem here is a little similar to that shown on apps.ouwu.com however it is simply down to a setup step I didn’t follow through correctly. Within the monitorix.conf file there is the following line:

our $OSTYPE="Linux-Debian";             # choose your OS type (see below)

Linux-Debian isn’t the default, and it needs to be changed to it otherwise none of the directories setup for your debian install will be used, it will instead default to something else. A little silly if you ask me but a quick fix. Upon restarting the monitorix process all works fine.

# /etc/init.d/monitorix.sh stop
Stopping Monitorix:
# /etc/init.d/monitorix.sh start
Starting Monitorix:  ... Ok

Now you have the process running fine, you will notice that the rrd files in /var/lib/monitorix/ are not being updated every minute or so, a quick fix for this as see in steps five to seven at apps.ouwu.com is to simply do the following:

mv /etc/cron.d/monitorix.sh /etc/cron.d/monitorix
chmod +x /etc/cron.d/monitorix
ls -l /var/lib/monitorix/

Repeating the last line a couple of times over the next few minutes should show that the files are being updated and you can take a sigh of relief. One last thing you may notice is that by going to http://your.domain/monitorix gives a 404, that could be because your domain is set within apache config to another directory within /var/www/ such as would be the case if you have a lot of domains hosted on that server my /var/www/ directory contains many sub directories each for one of my domains.Therefore you may want to set up a subdomain to forward to it, or simply set up a symlink pointing to /var/www/monitorix from your /var/www/domain.dir.

An interesting thing to note however is that ping latency usually, even when it appears that the connection is being restricted, often never reach higher than 80ms. Also while the download speed, which should be around 3Mbit (on an 8Mbit connection almost three miles from the exchange) drops to the 56k speeds its been doing recently, the upstream rate has stayed at a constant 550Kbit (See graph above). If there where a fault on the line or as several people have stated too much strain on the copper wire I would be seeing correlation between the upstream and the downstream, something that the above graph obviously shows as not happening. Please note that the spikes in speed are from tests ran during early morning 1am to 6am, speed tests where ran using the thinkbroadband.com speed test tool.

As I write this the connection speed appears to be ok and I’m getting a constant 3Mbit connection at the moment, if you google for any broadband package your bound to find some chaff in with the wheat and it could just be that eclipse had a bad few days. However I shall continue to monitor the connection and post a further update is one is needed.

Don’t go anywhere near Eclipse if you ever want to download. Doesn’t matter if you have the top end unlimited package. Eclipse throttle ALL downloads. – Vambo

At 3am I can get a good 375kbs but at any normal time of the day I’m lucky to get around 50kbs. avoid them like the plague. – Not a Happy Man

Internet access is a total lottery and download speeds are pathetic. – Gabby

They still can’t get their act together to send an accurate final bill. I do not recommend Eclipse Internet Broadband. Ever. - Afmspace

Phoned their IT department to try and sort out the problems and in my opinion they were useless. - mdyches

Even though an unlimited, 8MB connection is promised, I continuously find it hard to reach a speed of 512kb - chimp14uk

So it appears i’m not alone, a large collection of bad reviews all documenting similar problems of poor service and terrible support. Unfortunately the broadband is paid for by parents as i’m staying at theirs for the summer and there is nothing I can do other than weather it out as they appear to be quite happy with it even with mine and my sisters complaints. While at university I have used ask4 and virgin broadband, both delivered me full 20Mbit broadband with no problems, we had one disruption with the virgin broadband and within a couple of hours of our call it was fixed I hate to imagine how long its taken eclipse to fix outages proof of their slowness is just a google away. Anyway less of the ranting, i’m off to look for an effective broadband package for going back to uni and possibly my own mobile broadband package for when at home.