Server failed, took time to recover

Well, what can I say about modern hardware, it is built to fail and fail it will. When I had my old server that was built from multitude of parts collected from scrap, it did not fail to run for almost 10 years nonstop. This time when I bought specific “reliable” hardware for the new server, it did not even manage to run for one year. Out of all complex things that can fail, it was PSU, a.k.a., power supply unit.

SilverStone Strider Gold S series 750W, was my choice of reliable power supply, with 3 years of warranty. Little I knew how unreliable it would be in long-term, and fail in the most busy time of the year when I have way more important things to do then look for replacement. This picture describes exactly how I felt about the manufacturer of that PSU.

Anger
My feelings toward SilverStone

Once I managed to find a replacement (from different maker of course), I found that the “reliable” SSD managed to develop some unrecoverable bad blocks, where mysql (I am not a fan of this DB for its reliability) one of the most important files was stored. Boy, was I impressed with my luck …

As I have learned from long ago, if you do not back-up, you are guaranteed to lose data, glad that I had daily back-up that allowed me to restore database to the previous state.

One final thing to remember, always run your disks in RAID configuration, and make sure that it is mirrored to at least one more disk. Since my SSD was the only disk that was not in at least mirror configuration, it is the one disk that decided to fail, as expected.

Well, now is all in order, or so I hope, and it did not take as much time as I was fearing it would. Hey, and an excuse to make a post, though with no really useful information …

%d bloggers like this: