Fedora mirror at Faculty of Natural Sciences, Kragujevac, Serbia

Hey hey, I wish you all a happy new 2012. year. Do not be scared, Mayans predicted end of old and start of a new epoch(!), not armageddon (not even a bible one). How could they predict that??? All we know, is that they were a great mathematicians and astrologists. Some would say barbarians, but look at the conquistadors and their way of being civilized. Enough said. It’s time for some techy talk.

There comes a time, when every programmer wants to contribute to community somehow. Some do it often, some rarely or not at all. And if you didn’t, try it. It can be exciting. Remember, it is not the goal that is most important, but the way and the road to get there! Woah 😎 some wise words. Anyway, since I’m in the mood for wise talk, I’m gonna cut to the real deal 🙂 This is it:

Fedora mirror at the Faculty of Natural Sciences in Kragujevac, Serbia. Check it out if you are near it (geographically, I mean 🙂 ). There’s a image/link at the right side of the text.

The thing with supporting distribution by providing a mirror is that is useful, and not hard to do it at all. Well, of course if you fulfill the requirements. In this case it is some outbound bandwidth (100 MBit/s) and minimum 1 TB of storage, for all ISO and RPM packages. The procedure and all that “papers”, is not very big problem. As a matter of fact, it goes very fast. You can become a public mirror in a couple of hours. If you use a RPM distro as a system on a mirror machine, that is even simpler. In this case, Debian is used. But that is not a big problem. Downloading mirror-manager software from git repository, configuring a config file with a data you enter at Fedora Accounting System, and rsyncing all data, you can start to provide a mirroring services to all people from the world 🙂 Yey! I even installed a ganglia monitoring system, so I could watch how much it is used. Here’s a network traffic ganglia graph:

Enjoy the holidays!


Ganglia HDD plugin

Hey hey,

Like I said, it was lot going on. One of things was my first ganglia plugin. Let’s cut to the meat! 🙂

Plugin is developed for Debian like and Red Hat like systems (deb and rpm package). It is for personal use on grid cluster. Private in sense I did not write all documentation for packages, like for example man pages.

How I did this? I have used gmetric ganglia set tool, for making my custom metrics. Main tool is of course binary made by HD Sentinel, for listing hard disk info, like temperature, health, hours of work etc. Part of the package is script that is extracting basic data from HD Sentinel output and sending it to gmetad by gmetric script. To be precise, I used extended gmetric script written in Python. Reason was I needed grouping option, which default gmetric tool doesn’t have. GmetricP (P from Python) does grouping and sends it to gmetad daemon on main node with ganglia web page, displaying it under custom metrics.

Ganglia HDD Plugin

This is not very clean and pretty way of doing it, but it does the job. Better way would be to write ganglia plugin in Python, like described in ganglia documentation.