World Gone Web » Système

Debian Lenny: updating to SVN 1.6.4

samo — Wed, 04 Nov 2009 21:46:16 +0000

Working with Windwos users on my SVN repository, I needed to upgrade SVN to match with their Tortoise version. Here are the steps to do so:

add the line below to /etc/apt/sources.list
1
deb http://www.backports.org/debian lenny-backports main contrib non-free
get the public key from lenny-backports running the following command
1
wget -O - http://backports.org/debian/archive.key | apt-key add -
run aptitude update
install subversion aptitude -t lenny-backports install subversion

Debian: Naming USB devices with uDev

shalfon — Sun, 07 Jun 2009 16:49:16 +0000

I use three different small backup drives as three backup volumes. I turn them on on sunday, the backup scripts runs overnight, and I turn them off on monday morning. I’ve shared the backup script in a previous post.

The problem is that the backup script starts by mounting the backup drives. It expects them to have the same device name in /dev everytime. For example, it expects backup1 to be /dev/sde1. It turns out that these assigned device names can randomly change. Sometimes, backup1 appears as /dev/sdh1. The solution, in this case, is to assign a persistent name to each device. That is what udev is for. Udev will be called when a drive is plugged. If the plugged drive match a udev rule, it will be assigned the name defined in the rule.

The first step is to find criterias that will allow to identify a single drive and then create a rule for it.

Run the following command for the drives you need to name (replace sde1 by what you need):

1	udevinfo -a -p $(udevinfo -q path -n /dev/sde1)

In my case, I get the following result:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

looking at device '/block/sde/sde1':
KERNEL=="sde1"
SUBSYSTEM=="block"
DRIVER==""
ATTR{start}=="63"
ATTR{size}=="2930272002"
ATTR{stat}==" 48 1558 3506 228 19 0 38 248 0 156 476"looking at parent device '/block/sde':
KERNELS=="sde"
SUBSYSTEMS=="block"
DRIVERS==""
ATTRS{range}=="16"
ATTRS{removable}=="0"
ATTRS{size}=="2930277168"
ATTRS{capability}=="12"
ATTRS{stat}==" 60 1573 3722 504 19 0 38 248 0 348 752"looking at parent device '/devices/pci0000:00/0000:00:02.1/usb2/2-4/2-4:1.0/host4/target4:0:0/4:0:0:1':
KERNELS=="4:0:0:1"
SUBSYSTEMS=="scsi"
DRIVERS=="sd"
ATTRS{device_blocked}=="0"
ATTRS{type}=="0"
ATTRS{scsi_level}=="3"
ATTRS{vendor}=="ST315003" ATTRS{model}=="41AS "

1
2
3
4
5
6
7
8
9
10
11
12
13

ATTRS{rev}==" "
ATTRS{state}=="running"
ATTRS{timeout}=="30"
ATTRS{iocounterbits}=="32"
ATTRS{iorequest_cnt}=="0x57"
ATTRS{iodone_cnt}=="0x57"
ATTRS{ioerr_cnt}=="0x0"
ATTRS{modalias}=="scsi:t-0x00"
ATTRS{evt_media_change}=="0"
ATTRS{queue_depth}=="1"
ATTRS{queue_type}=="none"
ATTRS{max_sectors}=="240"
...

I have outlined the unique values and criterias that will allow me to identify backup1. I also outlined the subsystem lines that will be used in the udev rule.

The rule that will match this drive is:

1	SUBSYSTEM=="block", SUBSYSTEMS=="block", SUBSYSTEMS=="scsi", ATTRS{vendor}=="ST315003", ATTRS{model}=="41AaS", NAME="backup1"

More information about udev rules can be found online.

That done, you can test our rule by launching the following command:

1	udevadm test $(udevinfo -q path -n /dev/sde1) --force

If the rule matched the drive, you can ls /dev/disk/by-label and find backup1 in the list.

You can now edit the backup script and replace mount /dev/sde1 by mount /dev/disk/by-label/backup1. The latter should always refer to the right drive.

Thanks to the authors of the following two guides that lead me to this solution:

Debian: remove unused packages

shalfon — Wed, 15 Apr 2009 21:58:46 +0000

Two days ago, I ran into a strange problem with my email server which refused to send any more emails. It turned out my hard drive was full… I started by emptying the tmp directory and running aptitude clean and aptitude autoclean but that wasn’t enough to free enough space to allow me to breathe for long. Some Googling lead me to deborphan, a package that will list packages installed because they were once a dependancy but that are now useless to the system. Those packages identified, I just had to remove them all.

A few commands to do the same thing:

1	aptitude install deborphan

deborphan

1	aptitude purge <packages listed by deborphan>

Now my server can breathe and so can I.

Debian, read the bootup messages

shalfon — Wed, 15 Apr 2009 21:09:11 +0000

While Debian boots, several warning, error or information messages appear and quickly disappear from the screen. If you wish to log these messages in order to consult them later on, you need to enable the bootlog by editing a single configuration file.

1	vi /etc/default/bootlogd

The file should say something like

1	BOOTLOGD_ENABLE=No

Simply ,change it to

1	BOOTLOGD_ENABLE=Yes

You should now have these messages logged in /var/log/boot.

Access Subsonic via Apache

shalfon — Tue, 14 Apr 2009 18:03:45 +0000

I have setup Subsonic on my server last week as a standalone applicaiton. That means that Subsonic runs its own webserver. The only drawback of this method of deployment is that, to access the application, I had to type http://:8080.

To access Subsonic while using port 80, I enabled the mod_proxy and added a VirtualHost as follow.

To enable the proxy module, run :

1	a2enmod proxy_http

To access Subsonic via Apache, create a Virtual Host similar to the one below :

1
2
3
4
5
6
7
8
9
10
11
12

<VirtualHost *:80>
ServerAdmin <email>
ServerName subsonic.<subsonicserver>.com
ServerAlias subsonic
ServerSignature Off
<Proxy http://127.0.0.1:8080>
Allow from all
</Proxy>
ProxyPreserveHost on
ProxyPass / http://127.0.0.1:8080/
ProxyPassReverse / http://127.0.0.1:8080/
</VirtualHost>

Install Subsonic on Debian Etch

shalfon — Thu, 09 Apr 2009 12:06:19 +0000

Subsonic is a web application allowing you to stream or download your music collection over the Internet. It is the newest application of the kind after the famous Ampache and Jinzora. I have recently tried Ampache but was disapointed by the user experience so when Framasoft added a notice about Subsonic, I was immediatly interested.

The installation is pretty simple:

Download the standalone version
Untar the downloaded archive in the directory of your choice for example /var/subsonic
Edit subsonic.sh if you feel you need to change the default settings (port or memory allowed)

You are now ready to launch subsonic by running subsonic.sh

To use Subsonic, connect to http://:8080

Connect using the admin:admin account and you should now see the Subsonic main page.

To add music to your collection, browse to Music Folder and add your music collection, for example

Rock /mnt/data/music/rock

You can now log out, sign in using your user and enjoy your music from anywhere

Subsonic Main page

One nice feature offered by Subsonic is to reencode your music to a lower bitrate to limit the bandwidth usage, to do so, you need Lame. Download the lame archive from the official site, untar it, and rename the generated folder to “lame”. Then, copy this folder to /var/subsonic/transcode. You’re all done. Enjoy

Browse to the settings and immediately change the admin password. You can then create a new user for yourself.

Scanning for virus

shalfon — Fri, 24 Oct 2008 18:56:00 +0000

So now we have a backup system archiving our valuable data weekly. The problem is, these data come from different computers running different OS, some of them also come from friend’s computers or even the World Wide Web. The problem is that we can not guarantee that these files are virus free. But at least, we can scan them before we back them up. That’s why I created a script launching a virus scan every thursdays before we run the sunday backup. Not knowing which antivirus to use to detect windows-aiming virus I decided to go with Avast that I used when running Windows. Avast provides an Avast 4 Linux edition, free for home use.

The installation is straighforward: download the package from avast site then run:

1	dpkg -i avast....zip

The script is pretty simple too (feel free to help me improve it as well):

1
2
3
4
5
6
7
8
9
10
11
12

#!/bin/sh
export now=$(date +"%Y-%m-%d")
export logPath=/var/log/avast/avastlogs-${now}.log
echo "########## Starting Antivirus Script #########" > ${logPath}
date >> ${logPath}
avast-update
avast /mnt/media/ --report ${logPath}
echo "########## End of Antivirus Script ##########">> ${logPath}
export content=$(cat ${logPath})
mail -s "weekly antivirus scan completed" <email1>,<email2> << EOF
${content}
EOF

Then I added the task in cron:

1	0 1 * * 4 sh /root/config/cron/virusscan.sh > /dev/null

Hopefully that will save us some troubles…

Debian Etch and the Realtek ALC888 sounds chipset

shalfon — Sun, 08 Jun 2008 18:58:05 +0000

Until recently, my Etch based server was running without any sound driver. But I’ve heard about MPD and since that server also stores my music collection, I wanted to give it a try. Sound output then became necessary.

The mainboard is an Asrock ALIveNF6G-VSTA and, like the user manual and the Internet will tell you, the sound chip on that card is Realtek ALC 888. This chip is not recognized by default by Etch and requires a manual setup using a newer version of the alsa driver package. Here is the step by step installation procedure:

Download the latest alsa driver package.
Install the linux headers for your kernel

1	apt-get install linux-headers-$(uname -r)

Extract the alsa driver and cd into the directory, then run :

1	./configure --with-kernel=/usr/src/linux-headers-$(uname -r)

Run make (and be patient) then run make install

1 2	make make install

Run alsaconf and select your card (hda-intel) and follow the instructions on screen.

Now, launching the volume mixer of your choice, you should be able to set the sound and enjoy your speakers. You might also want to reboot to be on the safe side

I found this procedure on the Debian forum, so thanks to farrell2k who posted this solution on may 17.

CloneZilla: clone your hard drive

shalfon — Wed, 28 May 2008 21:28:53 +0000

About a year ago, when it was time for me to upgrade from Sarge to Etch, I decided to backup my whole Debian system before trying the risky aptitude dist-upgrade. I used g4u to clone my hard drive on another drive of the same size and even tested that, replacing the current drive with the clone, everything worked just as if nothing happened. It did and it also turned out I didn’t need that clone, the upgrade went fine.

Anyway, recently, I had to do the same manipulation for a neighbor’s (I fix the neighbors machines, I’m that way). The hard drive gave signs of weakness and I decided to just make the neighbor get a new drive and clone it. Somehow, on that machine, g4u didn’t work since I was not able to get a display going. Doing some research I found cloneZilla which provides the same cloning feature and is also a bootable CD.

The process is very simple and doesn’t require to read and type any command. One follows the GUI steps. First, decide whether you want to create an image of the drive or actually copy one drive or one partition to another. Then you select the source and the destination drives or partitions and press – it takes some courage – enter. The cloning took about 5 minutes for a 250Gb drive, and, again, replacing the aging drive with the newly cloned one, everything boot up and worked just like before. No data transfer, no windows setup, nothing…

To use cloneZilla, the simpliest way is to download the zipped iso, unzip it and burn it. It will create a bootable CD and you’re ready to go. It’s that easy so don’t hesitate cloning your drive if it starts to act funny.

Debian Lenny: install nvidia drivers

shalfon — Sat, 12 Apr 2008 23:02:14 +0000

Because my motherboard (ASUS P5K) and its SATA hard disk were not recognized by the Sarge installer I decided to install Debian testing aka Lenny. The installation was smooth (thanks to all the work done by the installer team) and most of my hardware was identified automatically (including for example the volume up/down buttons on my logitech keyboard). Unfortunately, at this point, I was using the nv driver and had no 3D support.

The installation of the nvidia driver should be pretty straight forward :

1	apt-get install module-assistant gcc nvidia-kernel-common m-a update m-a prepare m-a auto-install nvidia

This will only work after adding main and non-free to the sources.list for Aptitude…

Unfortunately, the installation is not always that simple and that’s why a debian wiki page lists no less than four methods to install the driver.

In my case, the problem was that I was using Lenny and that, at the moment, the package nvidia-kernel-source is ot available for testing. So I had to use a backport. Thanks to the help from the folks at on the debian irc channel, I found out how to do just that :

Add the following to sources.list and update the package list

deb-src http://ftp.debian.org/debian sid

1	main contrib non-free apt-get update

Build the missing package

1	apt-get build-dep nvidia; apt-get -b source nvidia

Install the resultant deb packages

1	dpkg -i &lt;packageName.deb&gt;

That done, the command m-a a-i nvidia finally worked. After restarting gdm, everything worked fine. Next step, test Beryl…