..:: seize the day ::..

It’s a name of an interesting book "Flirting With Disaster" by Marc Grestein and Michael Ellsberg written on 2008. But this is not about that book and of course I will not flirting with the disaster, it is just to remain me not to play a game with my current project . Well, I live in Indonesia. Here we live with natural disaster . We’re sitting on top of ring-of-fire, just two weeks ago, in early September 2009, I experienced 7.2 SR earthquake shaking my office in 17 th fl. building.

I was asked by French Red Cross to help them make Emergency Operation Centers in Bali, Yogyakarta and Jambi in Indonesia. In short they ask me to install several Linux server for their project. I will not tell you the detail design of the system, but I will share some interesting experience that I think could be useful to implement in your installation. The most important thing, considering the nature of our country, I should think to design of the high availability working environment for the system to be built. Moreover the center should be functional in the event of the crisis, whether it is earthquake, flood, electrical failure etc. Well it is a hard job, but I think it is the time for Linux to show the capability for managing such situation in my country.

There are several function in the Operation Center:

1. Web server
2. SMS gateway
3. Email and group ware server
4. IP PABX server
5. Database Server
6. GIS Server
7. Recorder and LDAP server 

I use several custom openSUSE 11.1 version that I build using SUSE Studio, thanks to SUSE Studio team that give the world the most wonderful re-mastering tool. We design the network on top of layer 2 HP switch with several VLAN, we also prepare the backup switch that already configure and when the main switch is broken the user can take off the switch and replace it with the backup.

We use Watchguard as router/firewall connected with two Internet connections that run on round robin high availability scheme. For the backup we prepare a Juniper that already configured that can replace the main router/firewall when it’s needed.

For the electricity backup we use a diesel generator that can supply enough power for the system, of course we use several UPS either for the server and for the workstation and other peripheral.

We design almost all the server in high availability with two servers running identical services in active-passive configuration. In implementing this, I consider the easiest implementation so the local support can follow our method and they can do minor troubleshooting by themselves. It is very important because I live far away from the sites. Of course I enabled remote connection using VPN and ssh but without local support it is not easy to do troubleshooting.

The implementation will be done in September - December 2009, so I will be very busy in the next 3 months. Nevertheless I plan to write my experience in several posting ahead.

Stay tuned and have a lot of fun.