Tired of waiting for 'resolving dependencies' on centos 3.x because of yum 2.0 ? Thanks to danieldk, yum-2.4.3 is available for centos 3 , as well as the newer yum-metadata-parser plugin. At the time of writing, it's only available in the [testing] repo, but you can enable it temporary and switch back to normal behaviour once that you've upgraded yum. BTW you'll have to clean up your /etc/yum.conf because newer yum will use CentOS-Base.repo sitting in /etc/yum.repos.d/ and so both centos repositories will be listed twice, which is not a real problem but a little bit annoying though ... My CentOS 3.x boxes now rocks .. :o)

Sun decided to release a free ODF plug in for M\$ Office users. Good thing ... usually OO.org users had to convert their OpenDocument files into either pdf (which is still the format i use for documents exchange with customers ...) or as a native M\$ office format. Now you can even send your document in odf format and they have no excuses anymore .... :o) . Btw it was also possible that such users install OpenOffice on Windows though ... Here it is : http://www.sun.com/software/star/openoffice/

A friend of mine (lefred) pointed me to another ldap administration tool but that uses a vi interface so useful tool in ssh connection for example ... it didn't exist in rpm for EL5 so i decided to build it and put it in my rpms repo ...

I was wondering how OpenMoko will fight against marketing around the iPhone from Apple .. i've now the answer : http://www.youtube.com/watch?v=POiNqP4savI .... :o)

I used several times z00dax's howto to remotely setup CentOS 4 on already installed Linux boxes (see http://www.karan.org/blog/index.php/2005/06/15 ) but if you want to test the same thing with CentOS 5 you'll have to add a new paramater : noipv6 . Otherwise anaconda will try to ask if you want an ipv6 address (the same behaviour that you'll have in manual setup)

Not a technical/linux tip, but as a consultant, i can assure you that such situation exist in the real life ... :o)

http://worsethanfailure.com/Articles/Im-Sure-You-Can-Deal.aspx 

I received a request from family members to be able to use Instant Messaging on a private server. I never installed a jabber daemon and if you have a look on http://www.jabber.org you'll see that a lot are existing ... Thanks to X-Tend's people advice i installed (very quickly) OpenFire . It exists already in a RPM format so i didn't have to build it. You can have your own jabberd running in 5 minutes and it has a lot of features, like different kind of authentication, different databases backend, etc .. it even has a plugin to act as a gateway for other IM systems, like IRC, AIM etc .... I was really impressed ... and the Admin console will make gui admins really happy :o)

I've received my new laptop (IBM Thinkpad R60) and this box has an ATI Radeon Mobility X1400 inside ... this card is not even usable by the xorg radeon driver. So i had to chose between using the generic (and included) vesa driver or installing fglrx (ATI driver). The second choice was the way to go it seems, but when you use fglrx, you can not activate 'desktop effects' because fglrx is incompatible with Aiglx (see http://support.ati.com/ics/support/default.asp?deptID=894&task=knowledge&questionID=26907 ) . But it's possible to use Beryl+fglrx through Xgl (and so not Aiglx) on CentOS 5. Xgl is not included but i decided to give it a try and i built it (with beryl as well) for i386 and i've included both rpms in my little repository : http://rpms.arrfab.net .

A lot of people misunderstand how NX/FreeNX works , especially the ssh authentication process ... How can you secure your NX box so that only key-based authentication will occur for the ssh part of it ? I've added some notes about it on the CentOS wiki here : http://wiki.centos.org/HowTos/FreeNX

You've added a scsi disk on a controller and you don't want to restart linux to scan the scsi bus ? Tell the kernel directly that a new scsi device was added .

The previous method was to use the echo command talking to /proc/scsi/scsi : For example , we've added a scsi disk on the first scsi controller (0), on the first channel (0), with a scsi id 4 (4) and on LUN 0 (0) : echo "scsi add-single-device 0 0 4 0" > /proc/scsi/scsi

Now (better and updated solution) : echo "- - -" > /sys/class/scsi_host/host0/scan

Check with cat /proc/scsi/scsi that you're able to see the device and use the disk without having to reboot ....