How are Hard Drives Binned and how to find good ones

[Perryg114]

I was wondering how hard drives are binned. I had my first major failure of a Western Digital Hard Drive last year after less than a year of use. Actually it still worked but had errors and would piss off my motherboard and I had to press F1 to get it to boot. It was some sort of smart drive failure issue. This was a data drive not a boot drive but still the bootup issue. The failed drive was a 640G blue series drive. Usually, I use black drives but I did not really know the difference. Other than price, how do you tell which drives are less likely to fail? I have had most of my black series drives out last the computer.

Manufactureres don't give detailed specs on MTBF and other important info. The WD green series drives seem to have reliability issues as well as the blue drives. I don't know much about Seagate drives and what different levels of reliability that they have depending on model. Are there any sub $100 2T drives out there that are worth a darn? I am talking something you would trust to put in a home server.

Perry

 

[taltamir]

1. All drives fail, no exception, keep backups. See the link in my signature.

2. With the exception of specific faulty models (which any company has), they are pretty similar in reliability.

3. SMART = Self Monitoring And Reporting Tools. you didn't have a "boot issue", the mobo was telling you "OMFG! The HDD just told me IT IS ABOUT TO FAIL! Back it up and replace ASAP!" and you pressed F1 to IGNORE that warning. Its pretty odd that you did since typically they EXPLAIN the situation clearly. SMART catches about 40% of drive failures, and not much before it happened (copy things off of it in order of importance when you first get it, there is a good chance it will die half way through transfer).

4. Green edition WD drives spin at ~5400 RPMs and automatically go to sleep after 8 seconds of inactivity. Blue edition @ 7200 RPM (requiring more expensive components to maintain reliability, making them more power hungry, & faster). Black edition are like blue, only with a beefed up controller (or dual controller) for extra speed.

5. MTBF on HDD is a BS marketing metric that means nothing of value or validity at all, ignore it. Yes, I KNOW it is calculated, I know how to calculate it too, the way it is applied in HDDs is meaningless... not to mention they fudge the numbers A LOT.

 

[pjkenned]

One thing to add to the above is that judging by internet reviews is a really bad methodology. WD Greens cost less than half of WD Blacks. So you will see more users with Greens and more "my hard drive died" musings on the internet with Greens than WD Blacks because of the numbers of drives that are out there.

Rule is, current big SATA drives: RAID 1/ 6. Even restoring 2TB+ of data over a network takes awhile.

 

[taltamir]

One thing to add to the above is that judging by internet reviews is a really bad methodology. WD Greens cost less than half of WD Blacks. So you will see more users with Greens and more "my hard drive died" musings on the internet with Greens than WD Blacks because of the numbers of drives that are out there.

yes, I was basing my "about the same failure rate" on the google study, the biggest scientific study of HDD failure ever made. with a ridiculous amount of drives of ALL makes

 

[WhoBeDaPlaya]

4. Green edition WD drives spin at ~5400 RPMs and automatically go to sleep after 8 seconds of inactivity. Blue edition @ 7200 RPM (requiring more expensive components to maintain reliability, making them more power hungry, & faster). Black edition are like blue, only with a beefed up controller (or dual controller) for extra speed.

Another thing to consider is the excessive number of parks for WD's Green drives. You could see the actual count in earlier models, then they wised up and just stopped reporting it.
Haven't had any Greens die yet (a 1TB, couple of 1.5TBs) *knock on wood*

 

[VirtualLarry]

Another thing to consider is the excessive number of parks for WD's Green drives. You could see the actual count in earlier models, then they wised up and just stopped reporting it.
Haven't had any Greens die yet (a 1TB, couple of 1.5TBs) *knock on wood*

They don't report head load/unloads in SMART anymore? Those bastards!

 

[taltamir]

there is no indication that the higher number of parks negatively affect reliability overall, just as the 7200 rpm does not do so.
Keep in mind that those things are accounted for in the construction and design of the drive. and that even if something does negatively affects reliability, it is still averaged out with all other reliability affecting factors to get overall reliability.

no drive is reliable so just keep backups.

 

[Perryg114]

I plan on keeping backups. I have a Windows home server and have it setup to make redundant copies on different drives. Most of the stuff on the server is in 3 places, the machine it came from and two copies on the server. Right now, the server has two WD 1TB drives in it. I want to add a couple more 2T drives. To give me 6TB total storage and maybe 3TB with some degree of redundancy. The server is getting full with the 2TB I have now.

Perry

 

[taltamir]

windows home server is junk that is prone to dataloss... MS tried to fix it in the next version, but it was not finished by the time it came out so they cut out that feature altogether.

Use something like freeNAS or nexenta.

 

[Perryg114]

What feature are you talking about in WHS? The duplicate folder thing? So far it seems to be working well but I have not had a HD crash to prove that out. There are redundant copies outside the server so I think I am ok.

Perry