Archive for October, 2011

Important meetings

October 30, 2011 Leave a comment

Sometimes I like to treat this blog as a personal one and write down things that might affect in fundamental ways our future… or at least my own future, potentially.

Anyway, I’m currently living in Peru and had a series of meeting over the last two months with the new Electronical Government Office (ONGEI). Can’t say much for now (don’t want to spoil it) but I was in a meeting with the Prime Minister who seemed to agree with my proposal (in the big lines) and told me “you’re not Peruvian, are you?” in a kind of funny way (I managed to get the Peruvian accent so much that my Spanish friends make fun of me now). I’d also like to think I’ve been able to bend a little the shape of the future of the One Laptop Per Child project in Peru (on the side of Good) through my contact with the office’s director, but that was mostly a side-effect of what I hope is coming.

On Tuesday this week, I met with MySQL and MariaDB’s founder/creator, Michael Widenius (see picture). I didn’t really expect to like the guy, but as it turns out I felt soooo much related between the experiences of what happened on the MySQL-MariaDB side and what happened on the D0keos/Chamilo side that I ended up passing a good time. Those two historical events (for the projects themselves) happened pretty much at the same time, as well.

Yannick Warnier (Chamilo) and Michael Widenius (MariaDB) in Lima

This put me in the usual sharing mood, and I thought about Latinux. Now Latinux is mainly a Latin-American group (it’s at the same time an association and a company and an open-source software start-ups cluster) founded by my new friend (or at least close acquaintance) Ricardo Strusberg (*not* the guy on the left, that’s Santiago Gonzales), with whom I met a series of times during the last 12 months, and with whom we’re mounting the Official Chamilo Teacher and Admin Certifications (first one of them is currently in a review process and should be out before our Chamilo Users Day Peru on the 18th of November, so people will be able to come around, go to the workshop and get certified the same day).

So… coming back to the sharing mood. As it turns out, MariaDB suffers the exact same problems Chamilo does right now (and probably LibreOffice, to some extent). It is better than the software from which it moved away, it has most of the influent people in it, but there is no widespread adoption yet, because the original software kept the trademark and is still luring people into them being the best. Furthermore, the job was initially done so well, with so many people involved in the original projects, that it’s difficult to out-rank them with simple good intentions. A considerable marketing effort is necessary to make people know we’ve changed name, and that we are now better.

One of the reason I’m so fond of that certification effort is that it will definitely give a boost to our project, to have the ability to get yourself certified and prove you are a true professional. If we do it, we’re definitely moving one big step forward in terms of promotion. And so could MariaDB. So I mentioned it to Michael, who immediately agreed this was a great idea, and as I’m writing this post (which is two weeks later), Ricardo has been discussing the details of such a certification and they will be launching it at the beginning of next year all over America and France.

This specific encounter (that might shape the future of MariaDB considerably) was casually made possible by the organizers of the event, who made it possible for us to meet. So thank you guys!

I can’t believe how things can happen so casually, but one thing is sure: if all of this would have been in another, more popular context, I would already have “casually” met with a bunch of the most popular IT people on the planet (in fact, I have a few practical examples in mind involving friends of mine). So, in a way, Peru is a great place, but there are still other (great) places to be out there.

Sugar Camp Lima 2011

October 20, 2011 Leave a comment

Escribo este artículo solo porque me doy cuenta que por ahora es muy difícil todavía encontrar información sobre el evento relacionado a la imagen (= el software) que se encuentra en las XO, las portátiles del proyecto One Laptop Per Child (Una Laptop Para Cada Niño, o OLPC) que se está organizando para los 18 y 19 de Noviembre 2011.

Entonces las respuestas a las preguntas más comunes son:

  • Qué es? : Se trata de un evento durante el cual se harán la mayor cantidad posible de esfuerzos para generar una imagen software y varias guías para poner en las pequeñas laptops verdes. Esta imagen trae muchas mejoras que hará que los niños puedan (1) usar su laptop más tiempo (ahorra la energía mejor), (2) tener más recursos didácticos en Español y quizás en Quechua y Aymara, también en zonas sin conectividad a Internet, (3) usar su laptop con más tranquilidad, ya que una gran cantidad de problemas técnicos han sido resueltos, (4) conocer más sobre la ciudadanía, la nutrición y una serie de temas que son relevantes para los niños de todos orígenes
  • Donde? : Se producirá en el Escuelab, que es un lugar de intercambio en el centro de Lima, a 200 pasos del centro de la plaza San Martin
  • Cuando?: Los días 18 (en la noche) y 19 (todo el día) de Noviembre del 2011
  • Quienes irán?: Ya tenemos confirmada la presencia de una cantidad importante de actores de áreas de software libre, de tecnologías, de defensa de la ciudadanía digital y del gobierno, así como personas que tienen esta increible oportunidad de conocer el Quechua o el Aymara. Pero lo más importante es que Usted, quien está sensibilizado a este gran problema de la educación infantil en Perú, esté presente para tomar contacto con nosotros y ayudarnos en este gran esfuerzo.

El Gobierno no tiene la responsabilidad de hacer estas cosas. Nosotros como ciudadanos elegimos el Gobierno, nosotros tenemos que mostrarle lo que esperamos de el. Ven y ayúdanos a mostrárselo!

BeezNest estará presente y apoyará como pueda en el evento, a través de su conocimiento y de su ánimo.

Creating a software RAID array on an already installed Ubuntu 11.04

October 18, 2011 5 comments

Let’s say you got confused by a misleading fake-RAID feature on an HP Blade server and you decided to ignore that the Ubuntu installer was telling you it found 2 disks while it was supposed (if it was actual hardware RAID) to be detecting only one. And let’s say you are lucky to have 3 disks, and you only one to use two as the RAID array (and they do not contain your operating system, i.e. the / partition). You might wonder: “And now what? S**** you, HP!” (that last bit is if you left panic get you, of course). Let’s also say that you had an installation of a large production web application on /var/www, which is what you wanted to be on RAID, but that, by an incredibly lucky turn of events, it is just small enough to fit on what’s left of your first disk…

Well fear no more! There’s a solution and I tested it for you…

First take a safe copy of what is on that partition that you wanted to be in RAID, and put it on the disk that will not change. Also, unmount that partition. If it was /var/www, use:

$ sudo umount /var/www

If the system says it’s busy, shutdown Apache and make sure no user is currently in a terminal in /var/www… (use lsof |grep /var/www for example).

If you happen to read this Ubuntu RAID guide between the lines, you’ll catch it talks about “mdadm”, a software to manage multi-disk arrays. As you know, installing software on Ubuntu (or Debian) is dead-easy:

sudo apt-get install mdadm

Then let’s say you wonder how to use it… As you know, getting documentation in English on Linux is dead easy:

man mdadm

Then you’re up for a lot of reading, unless you don’t really care about the details and you want to try it quick. This is what you would then do, considering your two unused disks (or “devices” for the geeks) are /dev/sdb and /dev/sdc (you can get an idea from “ls /dev/sd*” or “df” or “fdisk /dev/anything-you-can-think-about” :

$ sudo mdadm –create /dev/md1 –level=1 –raid-devices=2 /dev/sdb /dev/sdc

…where /dev/md1 is the new multi-disks device you’ll create, –level is the RAID type you want to use (1 for simple mirroring), and –raid-devices is the number of disks you will have inside your multi-disks device.

You can then confirm that you don’t care about the partitions it may find on these. Then launch

$ sudo mdadm /dev/md1

To check if that virtual disk exists. It should give you something like:

/dev/md1: 465.76GiB raid1 2 devices, 0 spares. Use mdadm –detail for more detail.

Now you’ve got a device, but that doesn’t give you a partition…You’ll have to create one with “fdisk”:

$ sudo fdisk /dev/md1

fdisk> n

fdisk> p

fdisk> 1

fdisk> <enter> for default

fdisk> <enter> for default

fdisk> w

But the partition is not formatted. As you might know, formatting a partition (let’s say in EXT4 because you like modern stuff moderately) on Debian/Ubuntu is dead-easy:

$ sudo mkfs.ext4 /dev/md1

You now have an EXT4 partition mounted on a RAID1 mirroring device. Almost done: you want to mount the partition in your file system. Let’s say as /var/www (that directory must exist and it’d rather be empty):

$ sudo mount /dev/md1 /var/www

Now, you want this mounting to happen on its own when you reboot, right? To do that, you need to update your /etc/fstab. If you had already something mounted there before, chances are you will find its line is already in /etc/fstab . For example, if you had one of the disks alone mounted as /var/www, you’ll have this kind of line:

UUID=244b687f-f8d9-4a48-986a-8a1a8a8d33bd /var/www        ext4    defaults        0       2

Right? Well, you’ll need to edit that line to change the device ID now… But how do you get that UUID? No problem, just launch:

$ ls -l /dev/disk/by-uuid

You’ll see your /dev/md1 listed there, along with its UUID. Just copy that UUID and replace the older one on the /var/www line of your /etc/fstab. Done. Now I recommend you reboot to check everything is mounting correctly, before you delete your safe backup file.

Well, that wasn’t too hard, was it? I actually believed I would have to reinstall the whole Ubuntu, but having a very slow connection for the update, I preferred not to. Seems like I won this round! Thanks to all the guys involved in these cool projects! You really made my day.

Inspiración para el desarrollo de Chamilo

Extracto de algo que escribí para Learning Review en el 2009. Ahora cambiamos de nombre y somos Chamilo, con la misma idea…

Simplicidad de uso es la fuerza mayor de Chamilo [ndlr: eraDokeos]. Al ser una plataforma de uso simple, permite que el usuario no se pierda dentro de la herramienta. Nos enfocamos a guiarlo directamente dentro de sus cursos, con pocas cosas para distraerlo fuera de ellos. Además, realizamos una herramienta de calidad, que permite, entre otras cosas, que los docentes puedan crear cursos rápidamente con un esfuerzo mínimo y un resultado máximo en términos de aprendizaje del alumno.

Consideramos que desarrollar esta plataforma de código abierto, es nuestro aporte en la construcción de un mundo mejor, al permitir a quienes no cuentan con capacidad financiera para hacer una inversión en un sistema de e-learning privado, acceder a esta plataforma con todos sus elementos, sin otro costo que el esfuerzo humano para instalarla y la voluntad de usarla.

Creemos que la mejora de todos pasa por la educación de cada uno, y que todos tenemos el derecho a una educación con las mejores herramientas. No sólo en cuanto a costos se refiere, sino desde un enfoque de accesibilidad para personas con dificultades. Creemos que cumplir con este objetivo mediante la metodología del software libre, nos permite brindar una plataforma que responde mejor a las necesidades de los docentes y alumnos.

Categories: Chamilo, Dokkeos, open-source, Spanish Tags: ,

Interesting educause video Moodle vs Blackboard

As always, the most interesting features here are the comments by users…

I guess that my own opinion is that Chamilo is more secure and more open, anyway :-)

%d bloggers like this: