Radeon IGP on HP nx9010

I just spent a few hours trying to find out what the problem is with my Debian. I just replaced Fedora Core 3 to Debian unstable on my notebook. It’s a HP nx9010, equipped with Radeon IGP:

0000:01:05.0 VGA compatible controller: ATI Technologies Inc Radeon IGP 340M

There was no video from noatun and xine. After some googling around I found that my DRI is not working. Great.

However, all is fine now. I found this great site http://www.nixnuts.net/files/README.txt

Basically what I have to do was add these to my /etc/apt/source.list:

deb http://www.nixnuts.net/files/ ./
deb-src http://www.nixnuts.net/files/ ./

and do this:

# apt-get -u install xlibmesa-gl1-dri-trunk

This fixes everything, even the annoying flickers on my screen on high resolutions. Now I can enjoy 1280×1024 on external monitor! 😀

Good job John Lightsey, thanks a bunch!

p/s – Somehow I find that Debian runs faster on my laptop compared to Fedora. 😐

Note: This effectively changes your X server to Xorg instead of XFree86.


For the last few months, our office building has been doing a lot of generator testing and sometimes these tests went out of schedule. We began to worry about the office servers, as constant power failures may cause data corruptions although we use reiserfs for all of the disks.

My colleague and I look for good UPSes, and found out that APC’s UPSes work very well for Linux machines. After some price and Linux research, we decided to buy some APC Back-UPS CS 650. These models are not intended to be used by high-end server products but since we only have some clone machines as servers, a couple of HP Proliant DL140, and a DELL PowerEdge 1425SC there shouldn’t be much problem.

We bought 7 units of them from Startec in Low Yat, for RM320 each. We took a few hours last Saturday to install them and rearrange our servers. 4 of them are placed inside the server rack:

4 UPSes - 3 x HP Proliant DL140, 1 x DELL PowerEdge 1425SC

On the server side, we install Debian’s package apcupsd, and on the monitoring server (with Apache) we installed apcupsd-cgi. To make it work, /etc/apcupsd/apcupsd.conf are edited with the following lines modified:


Our UPS came with USB cables, different cables requires different configuration line for CABLETYPE. Don’t worry the explanations in the configuration files are very good.

Next, a couple of tests were done with apcupsd to make sure the servers behave well in case of power failures. When we are satisfied, the monitoring part is easy enough – add the monitored hosts inside /etc/apcupsd/hosts.conf. Here’s a snapshot of the monitoring page:


Local APT Repository: Part 2

Final Steps on Server

What we did in part 1 was to sync the repository. Now we need to make sure that our APT Server can serve the files for use by client machines.

For this purpose, you can either choose to export the filesystem using NFS, or using Samba. I chose NFS. In the server’s /etc/exports:

THE_PATH_WHERE_YOU_SAVED (ro,all_squash,anonuid=65534,anongid=65534)

Please make sure that you have nfs-user-server installed, and run /etc/init.d/nfs-user-server restart

Mounting on Client Machines

Now, we need to configure the machines that are going to use this repository. On the machine’s /etc/fstab put in this line:

NAME_OR_IP_OF_APTSERVER:THE_PATH_WHERE_YOU_SAVED /mnt/debian-apt nfs defaults 0 0

Don’t forget to:

# mkdir /mnt/debian-apt

To mount it:

# mount /mnt/debian-apt

Next, we need to reconfigure your APT by editing /etc/apt/sources.list:

deb file:/mnt/debian-apt/ stable main contrib non-free

*If you like, you can also put in lines for your apt to use the testing or unstable dists.

Updating APT on Client Machines

To test:

# apt-get update

Don’t worry if not much info is displayed. The next time you want to upgrade or install packages, you should see that it needs to download 0B of packages, looks something like this:

2 upgraded, 6 newly installed, 0 to remove and 467 not upgraded.
Need to get 0B/2399kB of archives.

***Don’t forget to change the variables written using UPPERCASE above to the values corresponding to your installations!

Good luck! 😉

Local APT Repository: Part 1

If you have a bunch of Debian machines to be taken care of, you must have been using APT a lot. APT is incredible, but if you use it to update multiple machines, you’ll be bored to death waiting each and individual machine to download packages. What if I tell you there’s a way we can download once, use multiple times? That’d be great, wouldn’t it?


  • A machine with at least 20GB disk space (to be the repository server) – http://www.debian.org/mirror/size
  • rsync (to sync with the repository)
  • Port 873 opened in the firewall (for rsync to access Debian mirrors)

Preparing the server

First, you need to download the script http://www.debian.org/mirror/anonftpsync

Change the following variables in the script:

  • TO : put in the path where you want to save your repository
  • RSYNC_HOST : put the mirror closest to you (http://www.debian.org/mirror/list-full)
  • EXCLUDE : put in the architectures that are useless for you. In my case, I exclude all architecture except for i386

Make the directory you specified in TO variable above.

Next, test your script:

# sh anonftpsync

If you can see the files being downloaded in the file rsync.log generated by the script, it is working. Go have a coffee or a nap, as the first sync takes a lot of time 🙂

If you have verified that all files have been downloaded properly, save the script to /usr/local/sbin/anonftpsync

Don’t forget to:

# chmod 755 /usr/local/sbin/anonftpsync

Automating download

This is just the matter of asking cron to do the job for you, say at 6.30am everyday:

30 6 * * *              /usr/local/sbin/anonftpsync

You are done! Now your ‘APT Server’ will sync itself with the chosen mirror everyday. In the next part of this article we’ll configure a machine to utilize this server.