The Ramblings of Ken

I just discovered at work that a drive in one of our production servers failed about a month ago. The server kept humming along like nothing was wrong at all.

I had email reporting setup to alert our team when there is a failure. Apparently it isn't working anymore. I'll have to look into the alerting system.

With this post I just want to express my gratitude that RAID exists. The acronym RAID stands for Redundant Array of Independent Disks. It allows you to essentially combine many hard disk drives into one large drive. However, that isn't all. RAID has redundancy built in, if you pick the right configuration. In a RAID array your data is spread across all of the disks, which allows for all of the disks to work together for better performance. Parity data is also written to each disk. Parity data is the critical part. This type of data is really many calculations performed and stored on each drive. Parity data allows all of the data on one drive to be reconstructed should a drive fail.

I won't go into detail on how RAID exists in this post. There's plenty of places on the web that explain RAID. It will suffice to say that we configured our RAID array in RAID level 6. This essentially creates double parity across all of the drives. Double parity allows two drives to go dead and still have intact data.

We have a replacement drive coming tomorrow. I'll get that installed and then our array will be back in an optimal state. Right now it is basically doing calculations every time data is requested.

Our users don't even have any idea that a drive in our server has been offline for over a month. I'm really glad RAID exists.