And I'll drop this in here for the sake of adding more knowledge and experience to this thread:
So I've been running a RAID-0 that most of you know of for my OS/Programs drive. In addition to that I used to be running two separate RAID-1 (mirror) arrays using a total of four WD Caviar Black 1TB drives. That gave me a total storage space of 2TB because that's how two arrays with two HDD's in each array mirroring each other works.
I recently filled that 2TB of space and needed more, so instead of running out and spending money on more HDD's I just distributed the 2TB to other computers I have access to and combined the two RAID-1's into a single RAID-5. This gives me 3TB of storage and I still have data protection if one of the drives fails for some reason.
Read statistics:
I can't get write statistics because HD Tune only allows you to do a write benchmark when there are no partitions setup on the drive in question. I was eager to get my data back onto this new RAID-5 so I never performed the write test. However I can tell you this, a software RAID-5 is not the way to go. I only get about 13MB/s sustained write speeds for large file transfers, it bursts at around 110MB/s to start with but decreases to 13MB/s within 30 seconds.
In the future I won't both running RAID-5 or RAID-6 on a software RAID solution. There's a reason actual hardware RAID cards cost upwards of $200 when the support RAID-5 and RAID-6, they have their own processor to handle all the data striping and parity calculations. As a matter of fact I'm looking into setting up a 12TB RAID-6 using eight 2TB WD Caviar Black drives in the future here. I'm going to build a cheap system and then put in a quality hardware RAID card for the eight drives.
Bookmarks