The original PC stood on-line for 5 years without any maintenance or care (just automatic updates included in the distro), after 5 years the main disk got logically damaged, I simply replaced the disk reinstalled a newer distro, recovered the data from the old PC and voilà up again, during the change I got maybe 5 hours of no service at my Internet domain. That’s it.

Last year I got a newer PC and just replaced the old one.

I always wondered, why people spend so much money on real servers….?

Last week, because of the new implamementation of SAP, we got a server, an HP ProLiant ML350, that came as the recommend hardware by the SAP integrators. So I found it was time to explore my previouse question, this is what I found:

• As expected, it can be a very powerful machine, this one came with a 4 core Xeon processor, and 1Gb of RAM which we upgraded to 5.
• As also expected it has many expansion possibilities.
• Disks are removable and the system has RAID support built in.
• Everything is hot swapable, or at least swapable, therefore in case of any damage -like my previous disk damage- you got nothing or very few down time.

But, come on, a single hard disk for this server is more expensive that my current PC-Server! and it has more Gigs!

So my recommendation is, if you don’t run a mission critical service that doesn’t allow some downtime, just get a good PC and you can save some bucks.

Since Linux supports software RAID, just get that extra protection and you are done.