[Backblaze] - Hard Drive Reliability Stats for Q3 2015 - Page 7 - Overclock.net

Forum Jump: 
Reply
 
Thread Tools
post #61 of 88 Old 10-16-2015, 09:41 AM
 
aludka's Avatar
 
Join Date: Jun 2013
Posts: 397
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 37
Quote:
Originally Posted by iLeakStuff View Post

Cool. I store mine on NAND.
These clicking, slow, spinning disks should have died out ages ago along with Zip disks and floppy disks. Thats where they belong

I use good old fashion paper and pencil.....it takes me a while....but it hasn't failed me yet.

devil.gifCaseLabs Owners Clubdevil.gif
aludka is offline  
Sponsored Links
Advertisement
 
post #62 of 88 Old 10-16-2015, 09:57 AM
New to Overclock.net
 
Join Date: Feb 2011
Posts: 6,362
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 215
Quote:
Originally Posted by PR-Imagery View Post

Not the same logic, most of those drives that failed would work fine under conditions they're designed for. The only takeaway from the data is if you're trying to run a large 24/7 usage multiuser storage away some cheap drives are better than others.

but we're talking about probability of an abrupt failure.
in this case, any additional means of precaution against whatever would cause an abrupt failure simply implies that they're designed much better.

trolling an adult is very dangerous, don't try it at home nor at work. you don't want to play tag with a rabid man.
epic1337 is online now  
post #63 of 88 Old 10-16-2015, 10:33 AM
New to Overclock.net
 
Quantum Reality's Avatar
 
Join Date: Nov 2008
Posts: 6,349
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 308
I've been slowly replacing my older drives with Toshiba 2TB models from NCIX every time they diip down to $85 a drive, and I can definitely say I like having the elbow room. So there's always a place for electromechanical hard drives - you just can't beat them for large, cheap storage that is still reasonably fast compared to an SSD. smile.gif
Quantum Reality is offline  
Sponsored Links
Advertisement
 
post #64 of 88 Old 10-16-2015, 11:43 AM
 
misoonigiri's Avatar
 
Join Date: Jan 2012
Posts: 700
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 52
Quote:
Originally Posted by Quantum Reality View Post

I've been slowly replacing my older drives with Toshiba 2TB models from NCIX every time they diip down to $85 a drive, and I can definitely say I like having the elbow room. So there's always a place for electromechanical hard drives - you just can't beat them for large, cheap storage that is still reasonably fast compared to an SSD. smile.gif
Their 5TB is $139.99 at Amazon US, looks the lowest per TB at $28. But they seem to have 2 product lines for larger drives, with very different model numbers - confusing. 1 2 3

misoonigiri is online now  
post #65 of 88 Old 10-16-2015, 12:57 PM
Retired Staff
 
PR-Imagery's Avatar
 
Join Date: May 2011
Location: Dodging jet blast
Posts: 6,444
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 229
Quote:
Originally Posted by epic1337 View Post
 
Quote:
Originally Posted by PR-Imagery View Post

Not the same logic, most of those drives that failed would work fine under conditions they're designed for. The only takeaway from the data is if you're trying to run a large 24/7 usage multiuser storage away some cheap drives are better than others.

but we're talking about probability of an abrupt failure.
in this case, any additional means of precaution against whatever would cause an abrupt failure simply implies that they're designed much better.

 

using anything outside of it's designed parameters can lead to premature failure.


 
 
"The greatest good you can do for another is not just to share your riches but to reveal to him his own."

 

Project mATX

PR's Folding Stats

PR-Imagery is offline  
post #66 of 88 Old 10-16-2015, 01:07 PM
New to Overclock.net
 
Join Date: Feb 2011
Posts: 6,362
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 215
Quote:
Originally Posted by PR-Imagery View Post

using anything outside of it's designed parameters can lead to premature failure.

that just proves my point.

either seagate is doing an almost substandard design, barely meeting consumer standards.
or hitachi is just too damn good at making harddrives.

trolling an adult is very dangerous, don't try it at home nor at work. you don't want to play tag with a rabid man.
epic1337 is online now  
post #67 of 88 Old 10-16-2015, 01:39 PM
News Fiend
 
TheBloodEagle's Avatar
 
Join Date: Dec 2013
Posts: 1,397
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 45
A bit off topic but how do 2.5" HDD compare to the 3.5" HDDs in terms of reliability?

Corsair Graphite Club
Currently playing: War Thunder, Kerbal Space Program & Planetside 2
TheBloodEagle is offline  
post #68 of 88 Old 10-16-2015, 01:48 PM
Retired Staff
 
PR-Imagery's Avatar
 
Join Date: May 2011
Location: Dodging jet blast
Posts: 6,444
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 229
Quote:
Originally Posted by epic1337 View Post
 
Quote:
Originally Posted by PR-Imagery View Post

using anything outside of it's designed parameters can lead to premature failure.

that just proves my point.

either seagate is doing an almost substandard design, barely meeting consumer standards.
or hitachi is just too damn good at making harddrives.

 

But the data has nothing to do with operating under average consumer loads... so you can't really make that comparison.

 

And also

Quote:
Originally Posted by PR-Imagery View Post

Not the same logic, most of those drives that failed would work fine under conditions they're designed for. The only takeaway from the data is if you're trying to run a large 24/7 usage multiuser storage away some cheap drives are better than others.

 
 
"The greatest good you can do for another is not just to share your riches but to reveal to him his own."

 

Project mATX

PR's Folding Stats

PR-Imagery is offline  
post #69 of 88 Old 10-16-2015, 01:56 PM
 
Join Date: Jan 2011
Location: Brooklyn
Posts: 3,452
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 233
Quote:
Originally Posted by Chunky_Chimp View Post

Small sample size. The real TL;DR for anyone trying to knock Seagate for the nth time is not to bother with their odd-numbered-capacity drives. The writeup clearly points out that Seagate's 4TB and 6TB Barracudas are doing well.
Are you sure? Looking at the table it looks like they're including ~30,000 Seagate drives, which seems to be far more than any other manufacturer (HGST is around ~20,000 disks)...
Quote:
What I find confusing is why these guys only ever use and test consumer/desktop drives for a mission-critical environment, say they do that because of cost, and then let people draw conclusions on per-brand reliability based on tests of hardware NOT meant to last as long or run as well. A lot of people don't need enterprise hard drives, and probably won't ever have any, but again, this is talking per-brand reliability without accounting for the actual good stuff from each brand. I don't find that sensible at all.
It might be helpful to read their articles... they're pretty up-front about all of this...
Quote:
Originally Posted by ZealotKi11er View Post

Have 3 of those 3TB Seagates in my system. Difference is they are being cooled properly and they only act as storage.
Despite popular belief, running your disks cool actually decreases their life, while running them warm increases their life.

https://www.backblaze.com/blog/hard-drive-temperature-does-it-matter/
https://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/archive/disk_failures.pdf - Google's take
http://www.cs.virginia.edu/~gurumurthi/papers/acmtos13.pdf - MS's take
Quote:
Originally Posted by PR-Imagery View Post

they're actually pretty terribly designed and the fact Backblaze haven't shown vibration and shock test data (at least none that i've been able to find) all of the other data is irrelevant.
Just out of curiosity, how would they test this? I've never seen a DC provide this sort of information...
Quote:
Originally Posted by misoonigiri View Post

From further down the article, the failure rate is a percentage of the "Max # in Service" number (which may be smaller than the total # drives tested)
Quote:
Always consider the number of drives (Max # in Service) when looking at the failure rate. For example, the 1.5TB Seagate Barracuda Green drive has a failure rate of 130.9%, but that is based on only 51 drives. We tested these Seagate drives in one Storage Pod in our environment and they were not a good fit. In general, we’ve found it takes at least 6 Storage Pods (270 drives) worth of drives to get good sense of how a given drive will perform in our environment.
Repost for those who missed it.
Quote:
Originally Posted by Gilles3000 View Post

Indeed, I don't know whether their methodology has improved, but they used to have a lot of different enclosures. Some with proper cooling and vibration dampening, some without.

Hard to take their "testing" seriously with so many variables.
While the latter part is certainly true, they've been committed to their enclosures since 2013, and they actually have pretty decent dampening for the disks (source: I've built some).
Quote:
Originally Posted by XAslanX View Post

Yep, they are far from what I would consider a reliable source for information about failure rates. Their "pods" are very inconsistent from what a typical enterprise or home/office hard drive environment is like. Anywhere from improper/poor cooling to little to no vibration protection at all resulting in higher than normal failure rates.
While the data is not correlative of home/office environments, their pods are solid on cooling and decent on vibration.

Moreover, this is what a server environment's disk caddies look like:




It's not like they're suspending hundreds of drives at datacenters in suspended racks... they are basically the same as a home environment except, you know, much worse (having loads of racks filled with hardware in the same environment = lots of vibration).
Quote:
Here is a very good write up on why their methodology is bad http://bioteam.net/2011/08/why-you-should-never-build-a-backblaze-pod/
Did you read this write-up?

It says nothing about their test environment - at all. It's merely a warning to IT guys that their storage pods aren't practical for most data storage and are hard to maintain and therefore impractical for small offices... all of which makes a lot of sense when you consider that BB's model is get the cheapest disks and let them fail because we have backups (not exactly practical for most use scenarios).
Quote:
Originally Posted by ZealotKi11er View Post

Have you seen their system?
It is not much different from any other DC...
Quote:
The Data Center HDDs get read and written 24/7.
This is true in DCs, but not true in BB's environment - see misoonigiri's post above.
Quote:
Vibration also kills HDDs.
Was there a point to this?

BB uses better dampening than any consumer chassis I've ever purchased, and better than most server chassis... those rubber bands are thick.

Quote:
Originally Posted by Pandora51 View Post

The real issue here is they have different revisions of their enclousures. You don't know which drive is in a better enclousure and I doubt they mix up different drives.
This study is not 100% trustworthy. It can be useful and allows some conclusions but you can't simply be sure..
This is a great post.
Quote:
Since there is a completly different workload and propably a poor environment it will lead to high failure rates for seagate drives. But these drives could work very well in a consumer environment with low workloads. They are not designed for the server workload.
Now you might say they are worse than WD, Hitachi and toshiba drives because they are not able to work in a server grade environment and you might be right but what if seagates consumer drives have no vibration resistents due to cost savings?
With other words they would work completly fine for consumers.

But the last one here is wild speculation. Would love to hear an opinion from some of the HDD experts here.
Seagate drives lift their heads when vibrating too much.
Quote:
Originally Posted by PR-Imagery View Post

Not the same logic, most of those drives that failed would work fine under conditions they're designed for. The only takeaway from the data is if you're trying to run a large 24/7 usage multiuser storage away some cheap drives are better than others.
This post makes sense.
cq3mrd likes this.

I am a Marxist. Not the Sanders/Sweden/state bureacracy kind, but the workers' democratic control over the means of production kind.

I often use the debates I get into on these forums to demonstrate logical fallacies to to my students and debate teams I coach. Beware - your arguments are under scrutiny of adolescents and teenagers!
claes is offline  
post #70 of 88 Old 10-16-2015, 03:49 PM
New to Overclock.net
 
Blameless's Avatar
 
Join Date: Feb 2008
Posts: 29,121
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 3129
Quote:
Originally Posted by PR-Imagery View Post

Not the same logic, most of those drives that failed would work fine under conditions they're designed for.

Drives that survive beyond the conditions they are intended for will almost certainly last longer in the conditions they are intended for.

Any argument to the effect that more headroom or margin when it comes to durability is somehow not a good this a flawed one. These are all drives in the same segment with similar cost and similar performance...any statistically significant difference in failure rate is relevant if you want to have a drive most likely to survive in the long term.

My own modest sample size (about thirty drives at any given time and maybe a few hundred in total) closely mirrors Blackblaze's statistics.

...rightful liberty is unobstructed action according to our will within limits drawn around us by the equal rights of others. I do not add 'within the limits of the law,' because law is often but the tyrant's will, and always so when it violates the right of an individual. -- Thomas Jefferson
Blameless is online now  
Reply

Quick Reply

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off