Who Makes The Most Reliable Hard Drives?

I remember thinking of hard drives like this way back in the day (these are only a few familiar brands for each category):

LOW END
Matsushita, Wang, Teac, Mitsubishi, JVC, Fuji

MID-RANGE
Maxtor, Connor, Quantum, Samsung, Hitachi, Fujitsu

HIGH-END
Seagate, Western Digital, IBM, Toshiba

Ah, the good old days, those were simpler, better times in many ways.


And today, the list looks like this:

LOW END
...

MID-RANGE
Western Digital, Toshiba, Seagate

HIGH-END
...

As far as this list goes, it really isn't accurate, as each of these companies have HDDs in each category.

Low-end would feature WD Green and Seagate Barracuda LP and portable-class HDDs.
Mid-end would feature WD Blue/Black/Red/Purple and Seagate Barracuda/XT.
High-end would feature WD Velociraptor/Xe/Se/Re/etc and Seagate Constellation/ES/CS/etc, along with many other enterprise-grade drives.
 
^ Company in general. Many drives are only different in firmware, labels, and price tags.
 
So some oversensationalized nonsense notion amplified to extremes some 15 years ago regarding one specific line of drives from a manufacturer will stop you from buying a brand

Yep.

whereas the under-reported massive WD failures of the early 2000s did no such thing for you, the Seagate firmware issues with the 7200.11 series did no such thing for you, the Samsung SMART IDENTIFY data loss bug did so such thing for you

Haven't heard of any of that. So, nope.

I'll assume "never purchase anything Hitachi" doesn't mean other products they sell like tools, but hey, maybe an IBM product from over a decade ago will affect you buying another brand of unrelated products, I don't know.

Nah, wouldn't purchase Hitachi tools - prefer to stick with Dewalt.

What other media hype and sensationalism are you going to buy into next? Because it looks very effective for you, and I would like to make some money.

umadbro?
 
So some oversensationalized nonsense notion amplified to extremes some 15 years ago regarding one specific line of drives from a manufacturer will stop you from buying a brand

Actually, yes AND no.
Has nothing to do with oversensationalized nonsense, it has to do with the fact that 9 out of 10 hard drive related posts (at the time) were of people crying that their Hitachi Deskstars shit the bed.
 
^ Company in general. Many drives are only different in firmware, labels, and price tags.

Well, for the most part, there are quite a few differences between enterprise-grade HDDs and desktop-class HDDs.
WD Green vs Red, there is different firmware, yes, but there are also other RAFF-like features within the Red that are not present in the Green.

WD Blue vs Black, the Black has a dual-core controller among a few other features not present in the Blue.
If this were 2007, you would be spot on, but today, there are quite a few differences in the drives beyond what was stated.
 
Haven't heard of any of that. So, nope.

3Ll5IKJ.jpg


You, in a nutshell.
 
If consumer-grade drives are only slightly better than enterprise-grade, I will still take enterprise-grade drives over consumer-grade drives because (1) 5-year warranty bulldozes two or three year warranty to make up for it, (2) have a non-recoverable error rate of 1 in 10^15 instead of 1 in 10^13 (or 12) for consumer-grade, and (3) typically at least rated for 25% if not 50% more unload cycles.
 
The "Deathstar" nonsense is just as overblown as the Seagate Barracuda tick of death nonsense. That happened a very long time ago, and hasn't really returned. Though, 7200.x drives are still pretty unreliable "in my experience". I still have two "Deathstar" 120GXP 60GB disks in a Coppermine box that was retired after years of abuse, that I'd bet still work if I yanked it out of the garage and threw a CMOS battery in.

I don't trust new HGST drives, the manufacturing process change is too new to trust. I swore by Hitachi drives in production for multiple dual ZFS storage server iterations. After they were acquired, I moved to WD RE2 and had 3 failures in 48 disks in the first month at my scrub interval, then I moved to Samsung SpinPoint because I had them in desktops without issue, then after Samsung was acquired, I moved to Western Digital Red. It is important to me to be able to be to go to nearly any vendor and buy the same drive if I have to, years down the road. No surprises that way about geometry or weird firmware issues with my controllers.

Today? Starting over, I'd use nothing but Seagate Constellations for long term storage. It's the last refuge for consistent drive quality, really. Next years drive replacements are going to cost me a lot, but I might actually extend my paranoia of replacing long hour uptime disks to 4 years instead of 2 if I go that route...

Things I've learned about cycling two 24 disk raidz3 arrays every two years in bulk:

1. Don't trust OEM drives, always buy retail packaging. NewEgg and Amazon have given me the best shipping containers out of all of the vendors I've ordered from online (with the exception of NewEgg screwing up, years ago, but they've cleaned up their act)
2. Don't buy drives from Fry's in Renton -- they drop the boxes from the damned stair ladders if you ask them to check their stock because you need more disks than are on their shelf
3. Always keep your parity worth of storage for each vdev in cold spares near-by, unopened, because crap always goes down at 3 in the morning it seems like, and more failures can happen if it wasn't just bad luck
4. When a drive fails, RMA it, but never put the "refurbished" replacement back into your array -- give it to a family member or put it in a non-mission critical system and use it until it dies. Warranty is mostly irrelevant with drives, because once it dies, you CANNOT trust the replacement. Make the hard drive manufacturer send you a prepaid shipping label, too. (as this failure in warranty period just upped the drive TCO)
5. Always mount drives in something metal, don't rely on plastic rails -- direct heat dissipation is more important than airflow, even though airflow is good
6. Throw your power supply on a bench and scope it if you hear coil whine, or even if you don't, don't trust these enthusiast forums about power supply quality blindly. You will run into a bad sample, or the person reviewing it will be wrong. Even on the jonnyguru community.
7. Scrub every week, it will early fail weak drives well in advance to prevent worn "good" drives to fail at the same time as the weaker units
8. Hot spares are only useful in smaller arrays you can sacrifice, as a power supply issue may affect the hot spare the same time it does the rest of your drives...

Oh yeah, and finally... hard drives are going to die no matter what, it is not a matter of if, it is a matter of when, hence #3 :D

Two things to add. The new HGST NAS drives are in a brand new refurbished fab so it should have all the latest equipment and ISO specs, that is a good thing.

The deathstar fiasco was real and chronicled. The glass platters sounded like a good idea to increase speed but it also lead to degradation of the magnetic media which lead to read/write errors and bad blocks.

This hurt IBM enough to prompt selling the division to Hitachi.
 
My OCZ SSD died after one year of use. Seems really bad for something with no moving parts. On the other hand my WD 640GB harddrive is still being used now going on 6 years I think.
 
If consumer-grade drives are only slightly better than enterprise-grade, I will still take enterprise-grade drives over consumer-grade drives because (1) 5-year warranty bulldozes two or three year warranty to make up for it, (2) have a non-recoverable error rate of 1 in 10^15 instead of 1 in 10^13 (or 12) for consumer-grade, and (3) typically at least rated for 25% if not 50% more unload cycles.

Very well said.
I'm surprised no one else brought this up.

When it comes to many drives, consumer-grade is most likely better, just due to the cost of replacement.
But when it comes to fewer drives, I'd always choose enterprise-grade, especially for mission critical systems.
 
Enterprise-grade drives from Western Digital or Hitachi. Nothing beats enterprise-grade. Consumer-grade drives can just suck it!

I have a pile of dead WD RE3s and Seagate ES2s that beg to differ
 

I have a hard disk based audio recorder that still has a pair of 20Gb Quantum drives, they run great like what, 14~15 years later?... they made great drives (except the bigfoot, that one was terrible)
 
Western Digital has been the most reliable so far.

I would avoid Seagate at all costs, after that fiasco with their drives failing made me want to avoid them at all costs, I too had a piece of shit Seagate drive that died on me as well.
 
I have no enterprise experience, but for the last 5 years I have been running my own home NAS starting with 5 WD Greens, then growing to 6 WD greens, growing to 6 Greens + 2 Reds and now 12 WD RED 4TB drives.

In that time I have had only 1 drive fail a 2TB Green, but it was likely my own fault. To its defense it survived for 5 years while being mercilessly abused in one ZFS pool or another and I never patched it for the Intellipark issue, so it had some 500k head parks on it, when it gave up.

I've been happy with the reliability of WD Greens and Reds and probably will continue buying them.

It's funny, because I am usually not the "Brand Loyalty" type.
 
I've always run Maxtor and Seagate drives without issue, until I posted in a similar thread last year. Right after, I had two 7200.11 1TB failures; one was replaced under warranty, while I killed the other one during a botched BSY repair. I haven't turned away from Seagate yet, but another drive failure may change that (still have five running).
 
I've had failures in the first year on Maxtors, Western Digital, and Hitachi.
I had a Samsung last forever but the performance wasn't there.
I started paying extra for Seagate hdds and they've been worth every penny. Ive still got Seagate IDE drives going after 10 years.
That being said, I've got a Crucial ssd in my laptop and Samsung ssd in my gaming rig and both are fantastic. Haven't had either a year but oh dear god the performance increase is crazy awesome.
 
To lower manufacturing cost, most HD manufactures only secure the spindle on one side. This caused increased vibration and lowers reliability.

HDD are not really solid state machines. They are still very much analog and subject to analog errors. Moving mechanical parts are subject to movement error. They are semi quassi stable devices where if they fail to write or read, it will try again and again will it either passes or times out. (This happens more then you think)

To add to this, all drives are coated with a magnetic surface and this surface actually wears down with time because hard drive heads actually lightly glide ON the surface. So hard drives have a limited lifetime.

For this reason, if I have a lot of read only access, I go with solid state only now. If there is a lot of read/write access I use a RAID 5 array with WD blacks. I really don't find them that much more reliable then Seagates. But the WD RMA process is as painless as it gets.
 
WD black and RE based on blacks are attached on the top and bottom, this is featured. I don't see this mentioned with other manufacturers.
The stylus never touches the surface, it glides a micron or two just above it and with the latest Peizo to fine tune the read/write it's even more accurate.
If the stylus ever hits the platter surface that is generally the kiss of death when it breaks off. [head crash]
 
To add to this, all drives are coated with a magnetic surface and this surface actually wears down with time because hard drive heads actually lightly glide ON the surface. So hard drives have a limited lifetime.

No, they do not lightly glide on the surface.

If they actually touched the surface, the drive would be dead.
 
Hitachi is the winner this time around. Didn't see that coming. Very surprised by the higher than expected WD Red 3TB failure rates.

Guess I'll stick with Seagate Enterprise drives. Rock solid.
 
I have a pile of dead WD RE3s and Seagate ES2s that beg to differ

Same, got 2 Segate ES3 1TB 128MB cache drives, one DOA and the other died 1 month of use and does a very faint beep sound while trying to start, i hope the other 10 in production keep running while i RMA these...
 
So hard drives have a limited lifetime. For this reason, if I have a lot of read only access, I go with solid state only now.

SSDs are certainly already much more reliable than HDDs (the lack of moving parts is an insurmountable advantage) but SSDs have limited lifetimes too. It's just that SSD old-age failure is much more predictable than HDD old-age failure. You use HDDs until they fail of old age because there's really no guessing when they will fail of old age, but you cycle SSDs out before they fail of old age because you know when they will fail of old age.
 
Consumer drives shown to be more reliable than enterprise drives

http://www.computerworld.com/articl...-be-more-reliable-than-enterprise-drives.html
Some things to consider:
  • Enterprise drives spin faster
  • Enterprise drives are subjected to far more use
  • Enterprise drives are in environments where they are meticulously replaced as they have any issues. Consumer drives in consumer computers that die are likely to be "oh crap my computer died.. time to ewaste it :cough: throw it in the garbage :cough: and get a new one I guess."
 
Back
Top