Web Hosting Resource Kit – Editor’s Opinion 3/2004
Last weekend I received a call from a friend. One of his dedicated servers had crashed at 4.00 AM. The motherboard was fried and nothing worked anymore. This friend of mine is running a large auction website for art and partially depends on the income from this source. As the server went down his site was down. From the time he found the server down until the data center provider had fixed the box it took about 6 hours. I consider that pretty good. Anyway – this incident raised the question how much redundancy is needed for the average host? One provider can bring up that downed box in 6 hours but others might need 2 days. If your income depends on the server to be available 99.5% or more – can you risk running your server without failover or offsite/offserver backups?
Most web hosts offer some sort of backups but the number of RAID 1,5 server hosts seems to be limited and small. Is it time for the industry to reconsider the way it is heading right now? In most cases only people who get into a situation like this will look closer at redundant systems and failover technology. The additional money spend will not eliminate the chance for failure of hardware – but it will add an additional level of service availability to your business.
This friend of mine moved to a new server just a few days after this incident. Dual processors and RAID1 – if your income depends on the server you should not leave certain things up for failure. One downtime might be acceptable – imagine the box go down again because of a bad hard drive in 2 weeks. Clients will not always tolerate this. A good reputation could be destroyed in minutes.
So, what is your level of protection?