The long answer is theoretically, yes, but it’s not that simple. You have to look at this in terms of what it’s going to be doing. It’s going to be a fileserver not a gaming rig or even a PC for daily use. It will literally sit it the basement and dish out music and documents when I, or my family, need them. So any performance lost because I’m running from a live USB will go completely unnoticed.

I will have to disagree about FreeNAS being up-to-date though. They’re still rocking FreeBSD 8. The rest of the BSD world has moved onto 9.

I use it on my P8H77-I based NAS (LGA1155 socket, 6 SATA lines) with 8GB RAM, non ECC (small factor boards with ECC support and many SATA lines are hard to come by).

If you plan to leave it running 24/7 (mine doesn’t), ECC support is better, and you should be careful about the HDDs you choose.

To that end, I found this source most useful at the time (every 6 months the return rates of various components are analyzed), this is the last one: http://www.behardware.com/articles/881-6/components-returns-rates-7.html

The basic issue is that MySQL began in the space that NoSQL basically is in right now, namely light, fast dbs offering none of the guarantees you expect from an rdbms. At the time PostgreSQL wasn’t very mature, and had some issues, Oracle was a royal PITA, and the like. MySQL initially scored ease of use points and the fact that you could reasonably reliably prototype dbs on it which you couldn’t with PostgreSQL and Oracle at the time (no ALTER TABLE DROP COLUMN equivalent).

The two have moved in very different directions. PostgreSQL is moving at Oracle’s primary market, being an way to manage information where many applications may need to read/write the same relations. MySQL is moving in the area of offering something somewhere between a traditional RDBMS and a NoSQL db, continuing to offer fewer guarantees (since apps can set sql_mode) where only one app manages a given set of tables and all access goes through that app.