Bang! Yet another hard drive failure...

[Josh-H]

Within two months of my last hard drive failure I have had yet another - another Seagate 500 Gig Enterprise grade Barracuda from a 3 Drive RAID 5 set in my mac pro upped stumps and was booted from the RAID array for dying an untimely death last night.

That is the second Seagate 500 Gig Barracuda Enterprise drive to die in as many months.

And that makes a total of 5 hard drives that have died in 12 months for me including
2 x 1TB Maxtor drives - that were in a RAID 1 external case.
1 x 200 Gig Toshiba Laptop drive (that was thankfully a back up drive only)
2 x 500 Gig Seagate Enterprise Grade Barraccuda drives (Part of a RAID 5 array).

Currently rebuilding my mac pro in a new RAID10 array with 4 x Western Digital Black Caviar 1TB drives and restoring from Super Duper. My remaining Seagate Barracuda is being retired before it too cries *game over*.

All I can say is thank goodness for external back-ups.

I trust that is now the end of my bad luck with hard drives *touch wood*.

[budjames]

Wow! Bummer. I have my MacPro 8-core loaded with 3 x 1.5TB Seagate drives. I have 5 x 1TB Seagate drives in my eSata tower and 4 x 1TB Seagate drives in my ReadyNAS NV+. I upgraded my MacPro Laptop to a 500 Momentus 7400 rpm Seagate drive about 2 months ago. My external back up drives are all Seagate 1TB drives retired from my MacPro due to upgrades.

Knock on wood, I have never had a Seagate drive fail. I usually outgrow their capacity after a few years as larger drive become available.

I do have industrial grade UPS protection on all devices. If you do not, this may be a contributing factor.

Cheers.
Bud

[bill t.]

Am I just imagining this or do drives fail faster in raid systems than in non-raid systems?  Maybe it's just because there are a larger number of drives involved.  Or maybe the cabinets run hotter.  Or the drives are too tightly packed.  Or something.  Makes me wonder if raid is really so trustworthy.  After all, the drives usually share a single power supply.

I've had one drive failure in 8 years, which was probably caused because I failed to notice that my CPU was seriously overheating and it shut down during a backup-to-USB session.  The USB drive was OK, but the internal drive was scrambled.  Didn't lose anything.  Anyway I now copy new printing files to DVD as soon as I like them, in addition to once a week external drive backups.

For the last couple years have always bought Seagate non-enterprise drives, can not so far castigate them.  Except for the crappy Acronis software that comes with them, don't ever use it.

[bradleygibson]

Wow! Bummer. I have my MacPro 8-core loaded with 3 x 1.5TB Seagate drives. I have 5 x 1TB Seagate drives in my eSata tower and 4 x 1TB Seagate drives in my ReadyNAS NV+. I upgraded my MacPro Laptop to a 500 Momentus 7400 rpm Seagate drive about 2 months ago. My external back up drives are all Seagate 1TB drives retired from my MacPro due to upgrades.

Knock on wood, I have never had a Seagate drive fail. I usually outgrow their capacity after a few years as larger drive become available.

I do have industrial grade UPS protection on all devices. If you do not, this may be a contributing factor.

Cheers.
Bud

I managed to get four (of four) bad 1.5 TB Seagate drives and a friend I bought with got two (of two) bad ones.  This was last December, and all have been replaced by Seagate.

Last week one of the replacement 1.5TB drives (in an Areca RAID-6 array) went down.  Another replacement was shipped out and I'm back to double redundancy again.

That's a lot of bad drives--it's no stretch to say that the Seagate brand isn't my fave right now--this latest crop of drives isn't their best showing.

Josh, a number of colleagues have had issues with Maxtors as well over the past 3 years (again a small sample size).  My personal picks these days are 1) Western Dig, and 2) Hitachi.

[bill t.]

There was a known firmware issue with the 1.5TB drives at that time.  When I registered a drive online Seagate sent me an email to that effect, noting I did not have one of the problem drives and should not try to update the firmware which could apparently be done in the system via a download.  Moral...register your drives online.

[stewarthemley]

This sort of post scares me. It's why I asked if anyone was using blu-ray to store (http://luminous-landscape.com/forum/index.php?showtopic=36456).

Prices of discs are dropping fast: I'm in Japan at the moment and can get a 25gb disc for 198 yen. Less than $2. Getting more tempted by the day.

[Josh-H]

I do have industrial grade UPS protection on all devices. If you do not, this may be a contributing factor.

Yes - running on an UPS (also have very clean filtered power coming in anyway).

Josh, a number of colleagues have had issues with Maxtors as well over the past 3 years (again a small sample size). My personal picks these days are 1) Western Dig, and 2) Hitachi.

Yup - I have had bad experience with Maxtor drives as well. I am hoping these new WD Black Caviars are reliable (so far so good!) - The speed performance out of them is outstanding - particularly I/O.

Am I just imagining this or do drives fail faster in raid systems than in non-raid systems? Maybe it's just because there are a larger number of drives involved

I think its just the larger number of hard drives = more drives to possibly fail. In my case I think its just been a run of bad luck. All my drives are well ventilated and well looked after.

[bill t.]

FWIW my Inner Electronics Engineer tellsl me to check the power supply voltages on systems experiencing outrageous failure rates.  Both with and without the drive power cables connected.  Any major Geek'll tell ya.

[Justan]

FWIW, I work in IT and have done so for a long time.

Don’t ever expect a drive to last longer than the warranty. Plan for it to fail earlier. On mission critical equipment, it is a bad idea to ever leave a drive running (in service) when out of warranty.

It is a good idea to keep at least 3 regular backups of all data, with each backup set going to a different drive. Every night data should go to at least 1 redundant drive. It is a very good idea to also send it to another drive, intended for off site storage.

I'm not a fan of mirror sets as all they really do is to double the maintenance cost. In other words, when one drive of a mirror set fails, you are obligated to replace both drives. Also mirror sets don’t typically help in the case of inadvertent deletion. Worst of all, mirror sets still have a common failure point in the host adapter or anything else in the computer.

Best way to extend the service life of a drive is to keep the drive cool and keep it in as vibration free an environment as possible. you can cool a drive by putting a fan nearby or putting it into a cool room. My server rooms are all kept about 55 degrees (f) and doing this typically adds up to a year or 2 to the service life. Vibration dampening comes in a large part by the chassis in which the drive is mounted.

No drive manufacturer does any better a job at longevity than the others. Use the warranty as your guide.

In an average year I’ll replace about 20 to 30 drives, ranging in age from a few weeks to 3 years old.

Another suggestion is to let a new drive run for a couple of weeks before trusting it as a primary source for your data.

As the most recent example, I bought 2 2-TB WD drives. These were used half time and one failed 1 year and 1 month after it was placed in service, and the other failed 1 year and 2 months after it was placed in service. It is disgusting but that is why the manufacturer provides the warranty they do.

BTW, AMEX will double the manufacturer’s warranty, if you have the right AMEX card...

[woof75]

I've had almost every make of drive fail on me, I've had Lacie drives fail but I've never had a lacie rugged firewire drive fail and I've got 5 of them. They really do seem to be very good. If you look at the amazon reviews there's very few failures. (speaking of failures, I'm just trying to get a stupid maxtor to work and it won't, nothing important on it thank god)

[Plekto]

I've dealt with this for a very very long time as a computer consultant(plus various IT jobs).

It comes down to one factor.  Heat.

All drives need to have air flowing across them.  Often the drive is fine but the controller board or internal electronics die.  You drive(s) must have an intake fan blowing across them or be positioned with the air intake for the case over them(a simple way to do this is to put it in a 3.5 bay upside down and remove the 3.5 bay plastic panel cover above it.

It's also true for all drives that speed kills.  7200rpm is like 52x CDs - just really close to the limits of the media, 5400 rpm is far less hot and lasts a lot longer as well.  90% of people's cruddy performance with 5400rpm drives was due to their using on-board controllers(bad) or not using RAID.  A good way to get 5400rpm and low heat are laptop drives, which are great.  But not for RAID. (note - laptop drives ALSO need cooling, which is why so many external USB drives fail)

I use Western Digital drives only.   More specifically, their RE2 and RE3 drives.  These are optimized for RAID use/controllers and carry a 5 year warranty versus the regular 3 year for the other WD drives.  They are very reliable and as long as you keep them cool, they will run until their bearings die.  Using non-RAID drives for RAID often leads to dropped drives and other nastiness.   It's just not worth it considering the $10 or so extra cost per drive.

[Slough]

It sounds like there is some systemic problem with the drives. Having that many fail is not natural. In my experience a hard drive failure is rare assuming a correctly installed system. I've only seen a few go in 15 years in software engineering, working in quite large companies. At home I've had 4 or 5 blow, with three go within 6 months of each other, and I worked out that it was due to excessive heat. Don't assume that a product has adequate cooling just because it has slots for multiple drives. My Fujitsu Siemens PC, aka pile of poop, cooked itself. Sorry if this is obvious, but I really don't think you should have so many failures.

And yes your advice to double/triple back up is good.