Hey guys, I just saw a related thread by pincussion describing the imminent death of a hard drive and enigma's response intrigued me. http://www.xoohq.com/threads/33311.Sigh-Hard-drive-may-fail-soon. Rather than hijacking, I've got a new thread going. Problem: My server EATS hard drives at about one per every six months. I get maybe 18 months out of a hard drive. Question: Is this the cost of a sweet media server? Or, should I buy better hard drives? Or, the real question. Is there a way to under clock things like the chipset or other parts of the motherboard with the specific goal of increasing hard drive lifespan? Optimization software? Thoughts?
In my opinion, hard drive companies are feeling too much pressure for capacity and speed with SSDs breathing down their neck's. This is causing capacities to be pushed far too aggressively and new chips to be pushed out without enough testing. The QC has also gone down a LOT to save money so that prices can be lowered. All this leads to massive failure rates compared to older drives. The majority of users dont see this because they only have 1 or maybe 2 drives. But people like us who have 5+ drives and put the drives to more use see the effects to a much greater extent. I have owned from these companies in the past decade: Seagate: 4 drives Samsung HDD (Seagate owned): 4 drives Western Digital: 5 drives OCZ: 2 Samsung SSD: 7 Intel: 1 Of those, failure rates have been: Seagate: 1 drive Samsung HDD (Seagate owned): 3 drives Western Digital: 4 drives OCZ: 2 Samsung SSD: 1 Intel: 0 My next HDD purchase will include a Hitachi drive (owned by Western Digital) because I havent tried them yet. I hear bad reviews, but at this point Im willing to try anything for drive reliability. I am about to upgrade (because drives are wearing out right now and I need more space) to two 4TB drives from Hitachi and one 4TB drive from Western Digital. Seagate seems to be smart enough not to push 4TB before they are ready so none are available yet except in the enterprise market. These drive failures are not on a single motherboard, I went from an AMD 800 series board, to an Intel P67 series, and am back on an AMD 900 series now. Failures have spread equally on all boards. I ahve also used 2 different towers, each with different orientation on the drives. The Cooler Master had the drives sideways and I did see higher failure rate that way. This NZXT has normal horizontal orientation and drives still fail fast, but not as fast as when I was using the Cooler Master and had them mounted on their side.
You may be interested in this little tidbit Enigma. Recently had to go through wiping a ton of used 1tb - 2tb drives from Hitachi and Westerndigital both. In the end, had fewer damaged/fail wiped Hitachi's than Weastern Digital. Generally speaking on the average size drives, Western Digitals tend to still be useable when they come to us, while more Hitachi are failures. I thought it was rather odd after going through a good bit of the 1tb-2tb HDDs.
Thats weird, cause Western Digital owns and makes Hitachi drives. I guess its about the same weirdness as how "Samsung" drives made by Seagate are usually faster than Seagate's own branded drives, and for me how Samsung's have higher failure rates that Seagate branded ones. wtf?
WTF Indeed. Hey, think of it this way. They're -ALL- INFINITELY better than maxtor drives. . . I don't think I've seen a good one in a month. . .
Thanks, Enigma. So, you're saying its really an issue with the drive and there isnt really anything that can be done on the motherboard or in firmware/software to give it a bit of a break? Defragmenting isnt the issue. This is because I get a new drive in expectation of a failure every 6 to 12 months and make a backup once a month of any new files added in that month. I actually do this via a internal SATA/power connections I have fed outside the case. These extra drives literally sit in a little fire safe in my closet until its mate dies. Point is, the info that gets moved to it comes in gigantic chunks and is never deleted- no chance for getting fragged. Probably should have included that in the original post. Incidently Enigma, I run through a bunch of hard drives brands as well... Wanna know the funny thing? I agree with everything about your stats. My averages are almost identical. (When I adjust for the fact that on a two year time horizon the survival rate is next to zero.) I say "next to zero" because I have a Hitachi.... I know ... 500gb deskstar that was original a USB external box that I took apart to scavenge and use as the main operating system drive for the windows based server in question. All downloaded files go to it, first. Then, they are distributed to their respective homes on the storage drives and deleted. Same goes for any Rips. They go to this drive and then are sent in bulk to their final destination. Same goes for DVR backups. Live tv is recorded for two hours for rewind purposes and then deleted as time passes. To say this drive is hammered on a constant basis is an understatement. Whats even more, this was a drive that stored all my music and went back and forth to work with me before memory sticks were cheaply big enough. This alludes to my point. I bought this drive in 2009. It still works. The damned thing is a tank. long story, fascinating punchline.