Thank you for this. I'm looking for some reliable big drives for a NAS, this might be a good starting point.

awwwwww yesssssss. Love the data.

Wow, the one drive you least want to fail @18TB has a 12.5% failure rate! 10x worse than their other drives. Seagate ...

These charts always irk me because they don't take into account variables that can drastically improve or degrade a drive's reliability. For awhile, Western Digital Green drives were known to have high failure rates, but that was only because the 'head parking' was so damn aggressive (eight seconds; essentially a useless feature since hard drives park their heads while idle anyways). With a bit of firmware tweaking, you can disable it and essentially turn it into the equivalent of a consumer-level Red drive. This is also besides your typical statistical variances and deviations, ie. un-accountable spikes or dips due to small or un-uniform sample sizes. I'm sure there's SOME meaningful information that can be gleaned from this, but I feel majority of the information is a wash because there's so many damn un-accounted variables.

Mufflore:

Wow, the one drive you least want to fail @18TB has a 12.5% failure rate! 10x worse than their other drives. Seagate ...

Mo'e platters. mo'e problems.

Mufflore:

Wow, the one drive you least want to fail @18TB has a 12.5% failure rate! 10x worse than their other drives. Seagate ...

There's insufficient drive hours or count to draw any firm conclusions about overall reliability on that particular drive - 60 drives 2 failures.

tunaphish6:

These charts always irk me because they don't take into account variables that can drastically improve or degrade a drive's reliability. For awhile, Western Digital Green drives were known to have high failure rates, but that was only because the 'head parking' was so damn aggressive (eight seconds; essentially a useless feature since hard drives park their heads while idle anyways). With a bit of firmware tweaking, you can disable it and essentially turn it into the equivalent of a consumer-level Red drive. This is also besides your typical statistical variances and deviations, ie. un-accountable spikes or dips due to small or un-uniform sample sizes. I'm sure there's SOME meaningful information that can be gleaned from this, but I feel majority of the information is a wash because there's so many damn un-accounted variables.

Flashing a consumer drive to achieve reasonable reliability is imo an unreasonable expectation. It's impossible to account for all variables however thousands of drives installed in near identical pods, in a data center all running the same or similar software is about as close as we'll ever get outside of a lab. The biggest consideration is that AFR typically will increase with drive age and some drives this will spike substantially after a certain age point. Bearing that in mind however you definitely can glean some useful information such as how amazingly reliable the old HGST 4TB models are even if they're an older drive.

Mufflore:

Wow, the one drive you least want to fail @18TB has a 12.5% failure rate! 10x worse than their other drives. Seagate ...

The higher percentages are on drives that their numbers are small , only 60 drives the percentages are just not reliable since luck is the key factor. Now looking at the data darn the old hgst drives are workhorses !

tunaphish6:

. With a bit of firmware tweaking, you can disable it and essentially turn it into the equivalent of a consumer-level Red drive.

Care to share what tool one can use to tweak hard drive firmware. Am I taking as a computer tech who sometimes has data recovery tasks to solve where firmware is at fault.

i guess what most people look from the table is just the brand, which is more reliable overall if u look at those models, u guys should know that those (if-not-all) are enterprise hdds those what listed are WD ultrastar, seagate EXOs, and toshiba enterprise ... none such wd red,black ironwolf,barracuda do u guys using those hdds ? yes, probably enterprise and consumer not really that much different, especially in low-load enviroment probably there wont be big different, but in high-load+24/7 enviroment, in such case i think we can see the different if people want try, just use WD blue and black, black will last 2x blue most of time...based personal experience black is like enterprise level reliability for end-user

anticupidon:

Care to share what tool one can use to tweak hard drive firmware. Am I taking as a computer tech who sometimes has data recovery tasks to solve where firmware is at fault.

I take no responsibility for any individual's firmware tweaking shenanigans, but download and use Ultimate Boot Disk utility and use its WD Idle 3 utility to set the head parking delay/Intellipark (again, it's a redundant feature) to whatever you want, including fully disabling it. I would assume it's pretty straightforward from there, but in any case, you can Youtube specific instructions if you get lost. I almost forgot to mention that Green drives formally don't exist anymore, and have been re-branded under the Blue label--essentially any of their 5400 RPM Blue drives are actually their old Green drives. Mind, this was years ago when I last checked--who knows what Western Digital has been doing since, but in any case, I've yet to have a Blue/Green drive fail on me.

slyphnier:

black will last 2x blue most of time

While I can't really argue with personal experience, the reason why Green/Blue failure rates are so high is, by default, their head parking/Intellipark is set to 8 seconds. Conversely, their consumer-level Red drives are inherently the same drive, but with Intellipark completely disabled--hard drives crave consistency, and nothing could be more inconsistent than having the head 'park' every 8 seconds.

tunaphish6:

While I can't really argue with personal experience, the reason why Green/Blue failure rates are so high is, by default, their head parking/Intellipark is set to 8 seconds. Conversely, their consumer-level Red drives are inherently the same drive, but with Intellipark completely disabled--hard drives crave consistency, and nothing could be more inconsistent than having the head 'park' every 8 seconds.

I'll second this. My machine runs 24/7 handling house security. Since I turned off power saving features on my hard drives about 7+ years ago I only had one problem and that was caused by a processor overheating, causing a copy to corrupt about 4 years ago. The drive has continued running fine since. Before that I used to have one or 2 drive failures a year, it was disheartening. But I've been moving over to helium drives and those that arent helium are HGST, so that will have a bearing. Still, no drive failures in over 7 years now they are running 24/7 after my previous record is quite something.