I’ve had my Norco DS-1220 for six or seven months now and have some time to adjust to it and really feel out it’s performance. When I first setup the system (transitioning from the Thecus), I had a 2TB RAID 5 array built with 5 500GB Western Digital Caviars. When the array was at around 95% capacity I decided to add more storage and began looking for the most cost effective way to do that. I always use generic 3.5″ drives and I usually pay as close to $100 each for them as possible. Typically when the time comes for me to add storage it makes sense to double my capacity and Murhpy’s law always seems to work out that drives twice as large as the one I am replacing are at about $100 each.
This time around I paid a bit more then usual for the ‘WD10EACS‘ from Newegg. I bought the ‘Recertified’ version though which shaved about $20 per drive off the price and brought it to $109/drive.
Now, I’m all for ‘greening’ my life as it makes sense but I’ll admit I was not looking to buy specifically ‘green’ hard drives. As it happened though, the drives I bought were both the cheapest drives I could find and what Western Digital calls ‘Green Power’ drives. I was pleasantly surprised to find that WD had made drives both better for the environment and inexpensive. To sum up the technology in the drives very simply, the drives are able to spin at slower speeds when not in use to save energy and use about 1/3 energy of comparable drives. This decreased speed and power use equates both to the obvious reduction in power costs but also in a reduction in the heat generated by the drive. For me speed isn’t a major issue. The drives hold backups and media and are used by one or two people at a time in a non-mission critical application, but as it turns out I saw no decrease in speed speed compared to the ‘regular’ drives I am using.
As mentioned, I use the DS-1220 to hold my arrays. The DS-1220 is a simple box that holds SATA drives and presents them to the OS installed on the computer. I am not using a RAID controller and manage all of my arrays and using the software tools built into linux. Today, that means Fedora 9. The DS-1220 holds 12 drives and connects to the computer using 4 SATA cables. This is where things get a little hard to explain. The drives are not spread equally over the cables. Drives 1 through 5 in the DS-1220 connect to the machine via cable 1. Drives 6-10 connect via cable two. Drive 11 connects via cable 3 and Drive 12 connects via cable 4. Drives 1-5 and 6-10 are connected via one cable each through the use of SATA port multipliers built into the DS-1220 and are presented to the host computer as individual drives. If you were to plug 12 drives into the DS-1220 you will see 12 drives in the OS. The box is really quite dumb and by extension cheap and versatile.
Once I had the new array online I was able to do some simple, unscientific tests to see what the speed differences would be between the two arrays. The older 2TB array is built using WD 500GB drives and it not broken in anyway. While it’s capacity no longer meets my requirements as my main storage array, I am not taking it offline. I simply moved it’s data to the new array and then wiped it clean for holding other data. Since I had both array’s still online I thought I would check to see what the performance of writing a file over gigabit ethernet would be. The test was performed in a very basic, real world style. I took a large file, 7.27GBs and copied to both array’s over the network, one after the other. The network was in use and the arrays were both online and possibly in use during the tests.
The 4TB, Greenpower array copied the file in 5:41 seconds for an average speed of 21.32MB/s and it peaked at 38.73MB/s. The 2TB, standard, 7200RPM array copied the file in 5:38 seconds for an average speed of 21.50MB/s and a peak of 36.43MB/s. What does this show? Almost nothing. The bottleneck could have been the computer I was sending files to it from, it could have been the network or they could just be very evenly matched. It should also be noted that the new array sees more use then the old array simply because the older 2TB array now holds only backups and is not being hit with data requests constantly as the newer one is. I didn’t do anything to stop or slow those requests during the test so in the end perhaps the Greenpower, variable speed drives are in fact faster then the standard drives. Then again.. the standard drives are 2 years old so perhaps performance improvements have been made over all…
Whatever. Fact of the matter is the new, greener drives are perfectly capable of handling the load with no measurable performance hit in the real world. I’ve been running the array for several months now and have not had any drives fall out of the array or corrupt any data that I’ve seen. They are cheap to buy, cheaper to run and just as good as the regular stuff from a performance standpoint. It’s a good deal all around. I initally was worried that the drives would not be able to stand up to the work load of a RAID5 but so far so good.
More news as it happens…
Topslakr