I had a friend who wants a 100% secure and powerful servers. After doing the calculations on the load etc, I design a super redundant server. That guy can take big stress and traffic for a long long time. At least that is what I guess. All precautions taken, all drivers updates, whatever security enhanced. Well I thought that is safe enough.
Only one thing that is not very auspicious, he launched the site during Hungry Ghost Festival, but as techies we are not suppose to be superstitious. Well the server trial went fine and the launch was fine. It was running great all along. Two weeks later, red light emergency, one of the hard disk failed, that guy has at least a 5 year warranty super enterprise hard disk. Failed after 14 days. Really when something may happen, it really can happen. Ok I thought, thats nothing, the system was designed to be running even if two hard disks failed. It turned out kernel panic. The hardware can take it, but the Enterprise level software can’t take it. Took quite some work to get it back.
What else can I do? Out of 4 hard disks, 2 are already redundant. I guess I could have 5 hard disks with 1 hotspare or may 6 hard disks with 2 hotspares or………