Quote:
Originally posted by Wayne at Pelican Parts
Those are all good points, however I tend to think that most all of that computer equipment is not too well built these days. The way to get around this is to build multiple servers. We just bought new machines here for Pelican. We bought nine servers at about $3300 a piece. All with the same hardware configuration, so that if anything happened to one of the super-critical ones, then one of the not-so-critical ones could be scavenged for parts. All the disks are mirror 1 arrays too, so that if a machine takes a complete dump, we should be able to simply swap one drive into another box, and have it boot up. Theoretically. I don't want to have to test that theory.
-Wayne
|
All that like hardware, you should look into clustering, very nice redundancy in that scenario.