Hard Drive Experiences Please

[george__]

Having a really bad year -_-". I think consumer hard drives are going down the drain, quality wise....

Just this year...

Seagate Barracuda 750GB - dead but lasted me quite some time...
Western Digital 320GB Blue - I/O error (tried different SATA ports, different computers, and still giving error) (4 years)
Western Digital 320GB Green - clicking noise (so I'm thinking it's going to die) (4 years)
Western Digital 500GB Green - clicking noise (^^) (3 years)
Seagate 2TB x 4 - clicking noise (supposedly it's fine??. Firmware update didn't help) (1 year)
Hitachi 1TB - partition randomly wiped itself (but it's been working fine so far)

It seems like these newer drives are going to die even quicker than the older SATAs I bought :|.

What's other people's experiences?

Thanks!

[Mars2012]

I had a WD drive in a PC fail once and that's it. I have the Hitachi 1TB hard drive hooked up to the iMac as TM backup and it did something weird and needed to be reformatted, but since then has been working well. I tried to stay with Samsung or Hitachi drives, but it seems they've all been bought out by WD or Seagate.

I learned the hard way that as soon as you start hearing that clicking noise, back up all your info!!

I think a year or two is definitely too early for it to fail, but four or five years is pretty decent, especially if it has been written-over numerous times.

[Gee]

All drives fail. It is just a question of when

I stopped buying Segate after the 7200.11 firmware issue. But I still have Seagate drives in a few computers. I still have 2 IDE drives that have been running for 6 years

I have WD Green Drives in my NAS for the last 4 years and they are running 24/7 non stop
I also have Hitachi drives in my NAS for the last 3 years, also running non stop

No problems so far.

[Cafe_333]

Are all those drives in one pc? Have you considered overheating contributing to the failure of these drives? According to datarecoverylabs.com, the most common reason why hard drives fail is from overheating.

"A motor drives the spindle on which the platters are attached and rotates them at high speeds, ranging from 4,500 to 15,000 rotations per minutes (rpms). At this high rate the rotations generate heat. The faster the spindle spins, the faster data can be accessed, and the hotter the hard drive gets. As the temperature continues to rise, the platters expand, and then contract as the temperature decreases. These changes result in platters that change in shape and size. The result is a magnetic surface that is distorted and may have micro cracks and defects compromising the stored data." (link (http://www.datarecoverylabs.com/hard-drive-prevent-data-loss.html))

Perhaps if you improved the ventilation and airflow on to your hard drives it would minimize how much the platters expand and thus would not fail as easily. I have a fan directly blowing on all my drives in every pc even if it only has two drives. Hard drive cooling is very important for long-term reliability.

[ah802]

I think the reduction in warranty period pretty well sums up the manufacturers opinion on their own product. I've lost 4 drives under warranty in the last year.. 1 500g Seagate and 2 WD a 1.5TB and 320g, and a portable 260g the Seagate was a devastating loss, the O/S was on it... (no warning) along with 150 movies.. when it went, it held the computer from the bios. The three two gave signs... and left me hundreds of corrupt files. Back-up Back-up Back-up.... I found the best back up, is give the files to friends.. then you know were they are.. retrieval might be the cost of a dinner. Personal stuff save out to DVD's.. or some static solution.

If they're under warranty, then at least you will not be burdened with the financial cost. Some places have excellent extended warranties programs. Calculate a drive should last 5 years, anymore and you've had good luck...by the time you're ready for replacement the increasing size means you'll be able to put 3 drives on one for the same price, move the smaller drives out the door for whatever they bring (lest you be tempted to put less used valuable data on it) as a previous RFDer mentioned... ALL drives fail, just a question of time (like light bulbs). I understand Verbatim at Staples has a 2TB with 7 years warranty! RFDers should really be comparing warranties and letting manufacturers know that longevity is a priority in rotating storage.

Shopping for storage drives, I look at drives with less platters (means less physical parts moving), coolness of operation, quiet and smooth = long life, the ability to spin down (storage drives don't need speed, get and SSD if that's what you want) and no complex electronic features that could fail, and the best price and warranty.

I've got a couple of WD 80g drives still kicking over ten years old, they're in an old 'torrent' engine, but my recent experience shows those kind of drives were made in a different day. I look forward to 50TB drives, so I can put everything I've collected on one drive with room to space, then buy another for back-up.

[flyz]

General consensus is that the quality of consumer drives have gone down from anecdotal evidence over at [H] forums. Many of their user have 30+ TB builds and many many drives....

It seems the best bet is to look for drives with the longest warranty, and is why the new Seagate drives are hated because of the lottery warranty and general poor reliability and uptime (not built for 24/7 use). If you're going into a large build, have redundancy and some form of backup. It's also why a lot of [H] users go RAIDZ2 in ZFS.

They all seem to like Hitachi drives (though they're a part of WD...) for the quality compared to the other 2 brands. They're all quite excited for Toshiba built drives since they have the remnants of the Hitachi factory after the merger.

And this makes me very jelly. (also good resources in their HDD subforum)
http://hardforum.com/showthread.php?t=1708580

[arm2000]

A better way to look for HDD reliability is at retails sites with many reviews such as Newgg.com (the US Newegg). You will see that all HDD companies and all HDDs over 1GB have problems
Myself I had 2 2TB Hitachi HDDs last year that failed after just several days of use, one after the other. Using Seagate now.

[george__]

The drives I've mentioned weren't all in one computer.
They had fans blowing on them and coretemp showed them they averaged 35-40. Plus the pcs were in a very dark cool space...

I'm just freaken ticked I might have to face the idea of buying enterprise level hard drives due to the bloody unreliability of these consumer drives.
These newer 2TBs don't look like they'll last me even 3 years -_-"
PROBLEM is, enterpise level drives cost a FORTUNE

@arm2000 -be ready to go crazy due to the Seagate "clicking" noise

[arm2000]

There is no solid evidence that so called enterprise HDDs are any better, except if they are old models with lower density technology. The only way to guard against failures is to use RAID.

[george__]

There is no solid evidence that so called enterprise HDDs are any better, except if they are old models with lower density technology. The only way to guard against failures is to use RAID.

I disagree. RAID is probably stupidest way (but I'm moving towards Raid sooon). If you use raid, once you delete files it's gone and if raid goes crazy, you're spending hours working on fixing it.

Best way - make backups and keep them off site.

edit: Maybe if I can get some money, I'll get SAS :|

[flyz]

The drives I've mentioned weren't all in one computer.
They had fans blowing on them and coretemp showed them they averaged 35-40. Plus the pcs were in a very dark cool space...

I'm just freaken ticked I might have to face the idea of buying enterprise level hard drives due to the bloody unreliability of these consumer drives.
These newer 2TBs don't look like they'll last me even 3 years -_-"
PROBLEM is, enterpise level drives cost a FORTUNE

@arm2000 -be ready to go crazy due to the Seagate "clicking" noise

For the price of an enterprise drive, get 2 consumer drives and use one as backup.

[george__]

@flyz - That's what I'm doing right now.

For every drive, I buy a backup. But it seems these newer drives gonna die quicker ><. Freaking out

edit: I also don't like how warranty got so small now days.

[ugly]

I think everybody's experiences with hard drives are bad. And everybody has different experiences with different brands. I generally don't put much stock in what people say about hard drive brands unless there is an actual flaw with the product.

I have a 7 year old 200GB Seagate that has been rock-solid and is still in use. But most people say Seagate HDDs are garbage.

I bought a 1TB Hitachi about 3 years ago and it started getting bad sectors about a year ago. I replaced it with a 500GB Seagate and the Seagate has been fine.

Most people seem to rely on Western Digital hard-drives. Western Digital hard drives have been the worst for me. They seem to die every 3 years for me.

[arm2000]

I disagree. RAID is probably stupidest way
|
Huh?



edit: Maybe if I can get some money, I'll get SAS :|

What I/O interface has to do with reliability?

[mysticalinfluence]

A better way to look for HDD reliability is at retails sites with many reviews such as Newgg.com (the US Newegg). You will see that all HDD companies and all HDDs over 1GB have problems
Myself I had 2 2TB Hitachi HDDs last year that failed after just several days of use, one after the other. Using Seagate now.

:lol: :facepalm: You actually take Newegg reviews seriously? It's well known joke to never ever take advice from a Newegg review.

[george__]

SAS drives are designed for 24/7 operation.

RAID is stupid way to backup because...
a) if you accidentally delete a file, it's gone forever
b) if you use software raid, sometimes there can be issues :D.. I did anyway. Tried to truecrypt RAID partition (using the motherboard based SATA controller, it was a PITA).
c) but hardware raid controllers probably whole different story (I have a IBM M1015 but haven't tried RAID yet).

off-site backups of your drives you do weekly or bi-weekly probably safer bet. That's all. OR if you afford raid + backup.

I've had worst reliability with Seagate :D

[balance]

M1015 only does raid 0,1 without key attachment

People who get m1015 just want software raid like zfs

My 10x2tb is running nicely atm, but i'm hitting the gigabit bandwidth

[arm2000]

:lol: :facepalm: You actually take Newegg reviews seriously? It's well known joke to never ever take advice from a Newegg review.
I take seriously the sheer numbers of reviews on Newgg, much more than any other place. When you have such large numbers you can play statistics then.

[george__]

M1015 only does raid 0,1 without key attachment

People who get m1015 just want software raid like zfs

My 10x2tb is running nicely atm, but i'm hitting the gigabit bandwidth

Uh oh. I have one that i'm about to use... -_-" I'm going to get another one.
:D I went with SuperMicro + ECC memory route instead of Asrock route, which is why I needed the m1015. Too afraid of Asrock ><

<-- Guess I am sticking to Linux then :D. Save money not buying Windows Home Server or Server 2008 R2.

[arm2000]

SAS drives are designed for 24/7 operation.

So are all HDDs



RAID is stupid way to backup because...

You are only talking about non usual disk/RAID operations. None of them are even slightly good reasons to avoid RAID

[george__]

So are all HDDs


You are only talking about non usual disk/RAID operations. None of them are even slightly good reasons to avoid RAID

You win. I'm sorry

edit. I might get WD reds...
edit: I'm getting I/O errors now with another 500GB unit -_-".

Is keeping hard drives almost full (like 20GBs left) causing damage?

[arm2000]

You win. I'm sorry
Actually we all lose because this poor HDD reliability affects all of us. And regarding RAID I would say it's a rather unfortunate solution for this problem for the average computer user because it's rather cumbersome and never comes out-of-the-box.

[george__]

Actually we all lose because this poor HDD reliability affects all of us. And regarding RAID I would say it's a rather unfortunate solution for this problem for the average computer user because it's rather cumbersome and never comes out-of-the-box.

I'm scared of RAID. Imagine - someone comes and robs your home. ALL YOUR FILES on ONE COMPUTER. Too scary
I'm going to try ZFS :D. Balance will teach me!

[krazo]

I've lost quite a few WD drives in the last few years (including a USB external) due to bad sectors. Luckily I learned the importance of backups a long time ago (the hard way.)

These days I avoid WD and Seagate altogether and stick to Samsung.

I also run all new hard drives I purchase through full sector-by-sector read/write tests to ensure that it isn't already damaged.

[george__]

I've lost quite a few WD drives in the last few years (including a USB external) due to bad sectors. Luckily I learned the importance of backups a long time ago (the hard way.)

These days I avoid WD and Seagate altogether and stick to Samsung.

I also run all new hard drives I purchase through full sector-by-sector read/write tests to ensure that it isn't already damaged.

I have a WD that passes sector-sector test through hddtune but I still get I/O errors with 3 different pcs and a wack load of different cables. It makes no sense. It's the one I mentioned on my post

[balance]

Uh oh. I have one that i'm about to use... -_-" I'm going to get another one.
:D I went with SuperMicro + ECC memory route instead of Asrock route, which is why I needed the m1015. Too afraid of Asrock ><

<-- Guess I am sticking to Linux then :D. Save money not buying Windows Home Server or Server 2008 R2.

Going for an all in one :D

I need a 10gigabit nic and switch

My 1gigabit is my bottleneck

[george__]

Going for an all in one :D

I need a 10gigabit nic and switch

My 1gigabit is my bottleneck

jeesh. Can't do it where I live. (I think).
The RJ45 jacks are in the walls so I probably would need to renovate all the cabling in the home :|

[balance]

http://hardforum.com/showpost.php?p=1039003505&postcount=1696

My finish
Zfs build

[george__]

http://hardforum.com/showpost.php?p=1039003505&postcount=1696

My finish
Zfs build

Those drives gonna fry so tiny of a case

[death_hawk]

According to datarecoverylabs.com, the most common reason why hard drives fail is from overheating.


Google disagrees:
http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/archive/disk_failures.pdf

I disagree as well as I put HDDs into high temp situations all the time and I have a surprisingly low failure rate.
I had a couple 200GB Seagate IDE sitting in a plastic enclosure without a fan running at 65C with 24/7 read/writes going for 8 years. I only retired it recently.
Stupid drives were $1 per GB. But I loved them.

Even today my drives aren't exactly cool.
Most of my failures have been Samsungs. I had a DOA Hitachi once.
And I just had one of my 30+ WD Blacks complain and drop out of my RAID. I should pull it and see what it was complaining about.


I disagree. RAID is probably stupidest way (but I'm moving towards Raid sooon). If you use raid, once you delete files it's gone and if raid goes crazy, you're spending hours working on fixing it.

Best way - make backups and keep them off site.

edit: Maybe if I can get some money, I'll get SAS :|

RAID != Backups
The only thing RAID protects against is hard drive failure.



For the price of an enterprise drive, get 2 consumer drives and use one as backup.
Funny you should say that...
That's what I already do. Except not a 2:1 ratio.
I usually try to keep 10% cold spares, if not more depending on how long of a lifespan I'm expecting.

[george__]

Google disagrees:
http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/archive/disk_failures.pdf

I disagree as well as I put HDDs into high temp situations all the time and I have a surprisingly low failure rate.
I had a couple 200GB Seagate IDE sitting in a plastic enclosure without a fan running at 65C with 24/7 read/writes going for 8 years. I only retired it recently.
Stupid drives were $1 per GB. But I loved them.

Even today my drives aren't exactly cool.
Most of my failures have been Samsungs. I had a DOA Hitachi once.
And I just had one of my 30+ WD Blacks complain and drop out of my RAID. I should pull it and see what it was complaining about.



RAID != Backups
The only thing RAID protects against is hard drive failure.



Funny you should say that...
That's what I already do. Except not a 2:1 ratio.
I usually try to keep 10% cold spares, if not more depending on how long of a lifespan I'm expecting.

Just you wait, SATA drives are not comparable to IDE drive quality...

[balance]

Those drives gonna fry so tiny of a case

I did DD bench and copy 500GB data over

Smart info still reports back 29-33C

Going to check exhaust temp

I was surprise how much air it exhausted out

Might replace front 140mm with one that has more static pressure

Forgot to add a power saving cpu makes a big difference in power consumption

[death_hawk]

Just you wait, SATA drives are not comparable to IDE drive quality...

I still disagree.
I have over 100 SATA drives, roughly half in operation and still have very minimal failures.

As I said the only ones I have had failed are mostly Samsung, a Hitachi and a single WD.

[Mark77]

Going for an all in one :D
I need a 10gigabit nic and switch
My 1gigabit is my bottleneck

You can do LACP and aggregate the links. Intel NICs support LACP and those can be found on eBay for ~$40 a piece for dual interface. LACP supporting switches are cheap too. That should get your speeds into the 250mb/sec range over Ethernet without a lot of investment.

The big problem with 10gig-E right now is power consumption. It uses *a lot* of power to run all the electronics.

As for HDDs, been using Hitachi 7200rpm 2Tb's for the past 3 years, and very satisfied. Have a few 7200rpm Seagates that are about 5 years old, also still going strong. Only a fool doesn't use RAID these days.

[zod]

Over the years I've lost seagate, hitachi, maxtor, and WD hard drives. I actually lost two WD hard drives. One was past warantee, WD gladly replaced the other. These all died back in the IDE days. I did stop buying WD hard drives for a while. I ended up buying a 2tb green when the first came out (cheapest 2tb I could find to put in my popcorn hour a110 at the time). It is still going strong 3 years later. My computer also has a 1tb samsung spinpoint, 2tb WD Green, and a 1tb wd green, and an ozx agility 3 240gb drive.

The 1tb's are easily a few years old. The 2tb maybe two years old. I've been wanting to upgrade the 1tb's. I've been looking at more WD green's or a Hitachi drive to replace (the reviews I've read on the current seagate's leave me a little worried).

I suppose I've had all sorts of brands fail on me. My current batch is doing ok. I once held a gripe against WD, but this current batch is running great. I suppose I can't hold a specific grudge because I've experience failures from all the major hard drive manufacturers. (The 250gb IDE that WD sent me to replace a dead 120gb is still running inside of a modded xbox1.. lol).

[Cafe_333]

SAS drives are designed for 24/7 operation.

RAID is stupid way to backup because...
a) if you accidentally delete a file, it's gone forever
b) if you use software raid, sometimes there can be issues :D.. I did anyway. Tried to truecrypt RAID partition (using the motherboard based SATA controller, it was a PITA).
c) but hardware raid controllers probably whole different story (I have a IBM M1015 but haven't tried RAID yet).

I don't see anything wrong with running a Raid-1 setup as many businesses do. I also don't see how you cannot recover a deleted file from Raid-1 because it's technically a single drive that simply duplicates every file it writes to a second drive. As such I have on more than a few occasions, successfully recovered deleted files from Raid-1 setups. Although to be fair, i only use hardware raid - but i doubt that has anything to do with file recovery.

[Cafe_333]

Google disagrees:
http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/archive/disk_failures.pdf
I disagree as well...
If you had read my post thoroughly, it is not the high temperatures alone that's causing failures, it's the strain caused by the expansion and contraction of the platters. If your 200GB samsung was running 24/7 read/writes as you described it, this does not fit the profile i was describing. In any case, according to the link you quoted, Google seems to disagree with you.

"3.4 - Temperature is often quoted as the most important environmental factor affecting disk drive reliability. Previous studes have indicated that temperature deltas as low as 15C can nearly double disk drive failure rates. Here we report data on averages and note that other aggregation forms have shown similar trends and we therefore suggest the same conclusions."

So the next time you want to disagree with me, please cite why rather than making me spend time reading through a document only to find it contradicting your argument. The rest of the results in that document goes on to report that expected trends came back as expected, with the exception of low vs high utility which they found had little correlation with failure rate.

[george__]

@cafe_333 ... My point about RAID-1 is that if you don't have a backup, anything that gets deleted is gone forever. So it's like doing double work using RAID 1 because you're going to have to make backups anyway...

I think my problem is the drives are turning on and off so often because the files people access are are scattered all over various drives, so these drives turn on and off constantly.

[Cafe_333]

@cafe_333 ... My point about RAID-1 is that if you don't have a backup, anything that gets deleted is gone forever. So it's like doing double work using RAID 1 because you're going to have to make backups anyway...Isn't that what the recycle bin is for? If you accidentally deleted something, you can recover it right away via recycle bin. Unless you are talking about deleting something and then on the off chance you might change your mind about it months later, but i think that's kind of silly having a backup drive specifically for that reason... I suppose some people are just pack rats when it comes to files!

[balance]

You can do LACP and aggregate the links. Intel NICs support LACP and those can be found on eBay for ~$40 a piece for dual interface. LACP supporting switches are cheap too. That should get your speeds into the 250mb/sec range over Ethernet without a lot of investment.

The big problem with 10gig-E right now is power consumption. It uses *a lot* of power to run all the electronics.

As for HDDs, been using Hitachi 7200rpm 2Tb's for the past 3 years, and very satisfied. Have a few 7200rpm Seagates that are about 5 years old, also still going strong. Only a fool doesn't use RAID these days.

that's what I'm looking into atm

wonder if my netgear 3500L can support LACP switch, or I might but a 8 port gigabit switch which is pretty affordable for home use

going to have to look up how to enable LACP on my intel motherboard which have intel 82579lm vs 82574l via console on my OI/nappit NAS


upcoming windows 8/windows server 2012 is that they naitively support multiple NICs for single file copies (called SMB Multichannel), but would require a win8/ws2012 client and host :D~~~ can't wait for windows 8 support for trim over raid ssd and plus

[george__]

Server 2012 also going to have ReFS
<-- I'm a fool :( waaaa