Interesting data here on hard drives from Backblaze

[Toltepeceno]

I would say this is probably pretty good data as they have so many hard drives. I think HGST is on my short list alone, I guess they have shaken the "deathstar" name.

What is the Best Hard Drive?

At Backblaze, as of December 31, 2014, we had 41,213 disk drives spinning in our data center, storing all of the data for our unlimited backup service. That is up from 27,134 at the end of 2013. This year, most of the new drives are 4 TB drives, and a few are the new 6 TB drives.

Hard Drive Failure Rates for 2014
Let’s get right to the heart of the post. The table below shows the annual failure rate through the year 2014. Only models where we have 45 or more drives are shown. I chose 45 because that’s the number of drives in a Backblaze Storage Pod and it’s usually enough drives to start getting a meaningful failure rate if they’ve been running for a while.

*The table is a.t the link, the graph is below.*

 

[timothevs]

What clearly stands out is the low reliability of the Seagate 3TB drives. It is ignominious how Seagate produced such a shoddy product.

I had three of these drives fail on me in quick succession, a few months after the warranty was up. Never again am I going to trust my data or money to Seagate. I've made it my mission to dissuade people from buying their crap (pardon my French).

My seven year old Hitachi and five year old WD drives are still chugging along sans issues.

 

[Toltepeceno]

It's funny how times change. I have not bought any drives for several years and I had a couple of the old "deathstar's fail on me and several maxtor with a couple of western digitals. That's since the late 80's. I had not had bad luck with seagate. I'm glad to know this now as I will probably be in the marked to setup a backup system pretty soon.

Yes, the seagate 3tb info stands out like a sore thumb.

 

[timothevs]

I do remember the Deathstar debacle. But it seems that they recovered well from it. Luckily for me, I wasn't affected by that at the time, and any DeskStars I did have, I backed up, and unloaded on Craigslist quickly.

I've retired all my Seagate drives since I lost three in quick succession. They are/were really appealing at their price point, but I don't think I could deal with the loss in productivity due to another crashed HD.

That said, I do want to read up more about their new 8TB Backup drives, very enticing at $280 a pop.

 

[Toltepeceno]

I do remember the Deathstar debacle. But it seems that they recovered well from it. Luckily for me, I wasn't affected by that at the time, and any DeskStars I did have, I backed up, and unloaded on Craigslist quickly.

I've retired all my Seagate drives since I lost three in quick succession. They are/were really appealing at their price point, but I don't think I could deal with the loss in productivity due to another crashed HD.

That said, I do want to read up more about their new 8TB Backup drives, very enticing at $280 a pop.

Yep, 8tb a.t 280 is very nice.

 

[ColdCase]

I had three 4TB Hitachi drives go belly up last week (a 100% failure rate ), each at roughly 18 months. My 3TB Seagates and WDs are still jugging along.... I think there is some failure variances based on time of manufacture (batch) and warranty offered.... and even a wider variance between models of the same brand.

What model number are the drives, or are they a mix?

Seems like all my 2.5 inch portables have been solid even with the abuse they get.

 

[ZMacintosh]

interesting, as all of Apple's iMacs have either the 1tb or 3tb seagate hard drives.

i was looking at grabbing a few of the seagate 2tb or 3tb's to throw into my pegasus and i have their 6tb 5900rpm drive as my time capsule.

I was a huge proponent of Hitachi & Western Digital...but at this point, i don't think it really matters anymore.

 

[SaSaSushi]

Backblaze's last hard drive study last year was entirely discredited and it is amazing how soon people seem to forget.

This time around seems to be much of the same.

For what it's worth, I have owned several of the 3TB 7200.14 Seagate HDDs over the past several years and have yet to have any issues with them.

 

[Toltepeceno]

Taking tweaktown's critique seems more chancy than backblaze as they provide no data whatsoever to dispute it. Maybe it's them trying to get publicity. I do believe seagage has a higher failure rate a.t this point.

In fact here is a post a.t your link:

Bobby Cloutier · Prescott Valley, Arizona
Out of the ~500 drives (in serverrooms, in netapp/emc/tintri/tegile/zfs supermicros/delll servers) I work on (about 75/25 split of enterprise vs consumer drives) I can tell you that the overall trend is seagates fail ~3times as much for us, and to be honest on my desk now I have 24 good wd drives (to be added to servers after testing) and 5 bad segates (to be destroyed). wd greens all had to have there spindown time reset to make them usable thogh (and we no longer buy them because they do suck in raids).

There is also a person that said in this thread they had 3 fail. There is anecdotal evidence, like yours, saying the seagates fail higher. In fact I have seen a lot.

I have not had bad luck with seagate, but I have not bought a drive in some time and have never owned the 3tb. I do believe seagates quality has tanked and even if I take backblaze with a grain I also take tweaktown's article the same way, I have no reason to trust them any more than backblaze.

Backblaze's last hard drive study last year was entirely discredited and it is amazing how soon people seem to forget.

This time around seems to be much of the same.

For what it's worth, I have owned several of the 3TB 7200.14 Seagate HDDs over the past several years and have yet to have any issues with them.

 

[Toltepeceno]

Bingo. Tweaktown's critique has been entirely discredited. The debunkers have been debunked. It's amazing how soon people forget:

Trust Backblaze's drive reliability data?
Summary:TweakTown's tacit defense of industry silence on drive reliability - and against the Backblaze report - fails on multiple levels. Here's why.

http://www.zdnet.com/article/trust-backblazes-drive-reliability-data/

While my sentence is tongue in cheek above It seems that maybe it's tweaktown out for publicity, you can generally find things to go however you want on the internet.

 

[ColdCase]

Blackblaze's discredidation of Tweakown has also been soundly discredited by much more knowledgeable folks.

The best one can say about Blackblaze is it shows what may happen when you abuse the heck out of consumer level drives and use them in applications they are not suitable. The worst on can say is that its a publicity stunt and free publicity is good for BlackBlaze business, bad or good.

Its like running a Honda Civic or another economy transportation appliance at redline continuously around a race course trying to keep up with a LeMans car... The transportation appliances will fail, some sooner than later, but none will approach the LeMans car durability. Because they fail does not mean on is more or lest capable of years of mall cruising.. its completely irrelevant.

Some fan boys just don't get it.... sigh...

 

[Toltepeceno]

Blackblaze's discredidation of Tweakown has also been soundly discredited by much more knowledgeable folks.

It just shows you didn't even bother to read it, the discreditation was done by zdnet. Next.

Some fan boys just don't get it.... sigh...

Ah, name calling and on the first post of the thread even. The sign of one incapable of civil debate.

I don't know who backblaze or tweaktown is, but I do know backblaze posted data and tweaktown did not. You on the other hand posted an insult without even reading the article was by zdnet and not backblaze as you falsely claim (who i have heard of). I guess that makes you one of tweaktown. They usually skip any kind of debate and jump straight to insults just as you did.

 

[timothevs]

The Backblaze test may not be scientific, but they do subject all the drives in the study to similar conditions.

What does it say about a drive that fails while a counterpart (consumer drive - non NAS etc) chugs merrily along, under similar load/vibration conditions? I hardly think one can be called a "Fanboy" for merely observing that. To deny the facts, and to try and manipulate the observations on the other hand, does smell a wee bit like Fanboism to me.

Just my $0.02.

 

[Toltepeceno]

Thank you and I agree. I will gladly read any data to back up his claims, name calling does not qualify. I'm also not sure what I would be a fanboy of...actual data? In that case yes.

If one brand/model is failing more than others under ANY conditions it shows a problem with that brand/mode and as you say they are all under similar conditions. Not to mention data trumps analogies. A butt hurt seagate owner I bet.

Not to mention that those that fling around "fanboy" and "troll" the most fit the description best. I think thats my new sig.

The Backblaze test may not be scientific, but they do subject all the drives in the study to similar conditions.

What does it say about a drive that fails while a counterpart (consumer drive - non NAS etc) chugs merrily along, under similar load/vibration conditions? I hardly think one can be called a "Fanboy" for merely observing that. To deny the facts, and to try and manipulate the observations on the other hand, does smell a wee bit like Fanboism to me.

Just my $0.02.

 

[SaSaSushi]

While my sentence is tongue in cheek above It seems that maybe it's tweaktown out for publicity, you can generally find things to go however you want on the internet.

Backblaze's first study was widely criticized. Google it. I don't know if Tweaktown was among those who called them out on it or not. I was only pointing to the TT article as an example of criticism for their latest effort.

Blackblaze's discredidation of Tweakown has also been soundly discredited by much more knowledgeable folks.

Exactly, and considering this it's fairly remarkable that they're back at it.

The worst on can say is that its a publicity stunt and free publicity is good for BlackBlaze business, bad or good.

Well, I don't think even Toltepeceno could argue with the fact that they're name is getting quite a lot of press out of this.

 

[SaSaSushi]

What does it say about a drive that fails while a counterpart (consumer drive - non NAS etc) chugs merrily along, under similar load/vibration conditions?

If I may, I believe the point he was trying to make is that these drives are being subjected to load/vibration conditions that no average consumer (the Seagate ST3000DM001 is, after all, a consumer product) drive will ever come close to experiencing.

As mentioned, the ST3000DM001 has been a solid performer for me over the past few years. I have all my data backed up properly. They are sold here in Japan with a full 3-year warranty and for my level of usage, 3 years is about what I reasonably expect to get out of a HDD. I consider anything beyond that bonus time and never use drives older than that in data-critical situations.

But when the time comes, I'm probably going to start replacing my current batch with the Seagate 4TBs as the price per GB is improving daily.

 

[Toltepeceno]

Well, I don't think even Toltepeceno could argue with the fact that they're name is getting quite a lot of press out of this.

Backblaze's first study was widely criticized. Google it. I don't know if Tweaktown was among those who called them out on it or not. I was only pointing to the TT article as an example of criticism for their latest effort.

I did google it, early on. It was tweaktown that called them out and it was zdnet that called tweaktown out. I had already posted that, I don't expect seagate employees to read soemthing like that.

If I may, I believe the point he was trying to make is that these drives are being subjected to load/vibration conditions that no average consumer (the Seagate ST3000DM001 is, after all, a consumer product) drive will ever come close to experiencing.

It does show that some brands go under quicker under the same conditions. Even you can't argue with that.

 

[paulrbeers]

If I may, I believe the point he was trying to make is that these drives are being subjected to load/vibration conditions that no average consumer (the Seagate ST3000DM001 is, after all, a consumer product) drive will ever come close to experiencing.

As mentioned, the ST3000DM001 has been a solid performer for me over the past few years. I have all my data backed up properly. They are sold here in Japan with a full 3-year warranty and for my level of usage, 3 years is about what I reasonably expect to get out of a HDD. I consider anything beyond that bonus time and never use drives older than that in data-critical situations.

But when the time comes, I'm probably going to start replacing my current batch with the Seagate 4TBs as the price per GB is improving daily.

I have (had) 6 3tb Seagates, have 8 1.5Tb Seagates, 4 3TB Toshibas and 4 2TB Samsungs. All are in the same enclosures (mediasonic 4bay usb 3.0) and the only drives I have had fail are 2 of the Seagate 3tb at right at 1 year of service. The seagate 1.5tb drives are like 4 years old! While my numbers are not statistically significant vs the amount of drives sold by Seagate it is interesting that my own experiences mirror Backblaze's. I'm actually thinking about removing my 4x3tb array and replacing them simply because I have no faith in the remaining 4 drives. With that said I probably will go with Seagate 4tb because of their recommendations and because my old 1.5tb drives keep humming along....

 

[SaSaSushi]

I have (had) 6 3tb Seagates, have 8 1.5Tb Seagates, 4 3TB Toshibas and 4 2TB Samsungs. All are in the same enclosures (mediasonic 4bay usb 3.0) and the only drives I have had fail are 2 of the Seagate 3tb at right at 1 year of service.

Interesting and thanks for the feedback. Out of curiosity, were the two 3TB drives bought at the same time (e.g. same batch).

I was not questioning the bad experiences of anyone here, just sharing my own experiences and aside from having been a long time satisfied customer, I have no connections to Seagate whatsoever. I have had a few Seagate drives fail. For example, the 1TB drive in my previous iMac crashed. I've just never had a 3TB fail on me (yet).

 

[paulrbeers]

Interesting and thanks for the feedback. Out of curiosity, were the two 3TB drives bought at the same time (e.g. same batch).

I was not questioning the bad experiences of anyone here, just sharing my own experiences and aside from having been a long time satisfied customer, I have no connections to Seagate whatsoever. I have had a few Seagate drives fail. For example, the 1TB drive in my previous iMac crashed. I've just never had a 3TB fail on me (yet).

All 6 were bought in the latter half of 2013. I'm happy with Seagate drives for the most part. Just a bad model I guess. I had Several of the "Death Star" hard drives so I know how every company puts out a bad round of drives now and then.

 

[matreya]

I had Several of the "Death Star" hard drives so I know how every company puts out a bad round of drives now and then.

The DeathStars were IBM drives BEFORE Hitachi took over their drive division...

 

[paulrbeers]

The DeathStars were IBM drives BEFORE Hitachi took over their drive division...

Point being? Technically Hitachi drives are WD drives now, but you would still say the technology is based on the Hitachi drives right? Therefore when hitachi took over for IBM, they would have continued to build on the technology they purchased from IBM right? heck how many of the same engineers are now working for WD, that worked for IBM?!?!

Point is, all companies have a bad model now and again.

As for all the bantering about whether Backblaze data is valuable to a Consumer since they are using consumer drives in ways they weren't intended, quite honestly this is as close as we can get to a real drive comparison at a grand scale right? Just in this thread alone you have one guy who states his 3TB Seagates are great and I had the opposite experience. It isn't until you get to the 1000s of drives and their longevity that you can really determine whether one drive is better than another.

Like the Deathstars back in the day, from a purely benchmarking standpoint they were amazing. Longevity they sucked hard.... Unfortunately most websites only benchmark a drive for a few days to get the hard numbers. It is too costly to try to do longevity benchmarking (and frankly might not get those numbers for YEARS).

So let's all just agree that the data is valuable to at least give us all an idea of which drives can take a pounding and at least that should give you some idea of which drives to you should buy over others.

 

[VirtualRain]

I've been using HDs long enough to see all brands and sizes fail. My philosophy is simple... buy the cheapest high capacity drives available. 5 years ago that was WD Greens. Last year it was Seagate 4 and 5TB drives.

It's just not worth trying to game the odds by paying more for what you think may be a longer lasting drive. Buy cheap and replace as needed. (Oh, and make sure you backup)

 

[ColdCase]

So let's all just agree that the data is valuable to at least give us all an idea of which drives can take a pounding and at least that should give you some idea of which drives to you should buy over others.

Having much hands on R&M and systems engineering background and much experience using components beyond there design specs, I cannot agree with this.

Flawed data or trying to apply irrelevant data does not make the results useful, perhaps interesting. It is well known that using desktop type drives in a multi bay enclosures is a crap shoot as some durable drives are stressed by vibration harmonics, power variations, heat loads more than others. This data seems to support that well known fact for the drive enclosures being used. You may get different results with another enclosure.

I think we can all agree the vendor is getting lots of free advertising.... which is clever.

 

[paulrbeers]

Having much hands on R&M and systems engineering background and much experience using components beyond there design specs, I cannot agree with this.

Flawed data or trying to apply irrelevant data does not make the results useful, perhaps interesting. It is well known that using desktop type drives in a multi bay enclosures is a crap shoot as some durable drives are stressed by vibration harmonics, power variations, heat loads more than others. This data seems to support that well known fact for the drive enclosures being used. You may get different results with another enclosure.

I think we can all agree the vendor is getting lots of free advertising.... which is clever.

But many of us use this very approach and for us this is invaluable.