HDD failure rates

[nattivillin]

One of the biggest customer service issues we have had for the past 5 years are bad hard drives. Either in our consumer or business divisions, the failure rate is just simply unacceptable.

Of all the things we sell, hard drive returns are 10x any other part. We switched from Maxtor, to seagate, to IBm, to hitachi, and now we only sell Western Digital's. Still have the same issue. ( Slightly less with WD's) We figured we would pay slightly more for better hard drives if it decreased the number of unhappy customers. That has not worked. The more we sell, the more returns we have.

My question is, has anyone else seen this? Are there HDD's that have a higher reliability? Are we just ordering junk? Is it our supplier?

[natrat]

I'm seeing this too. I'm sure in the past 18 months hard drives have become hugely more unreliable. i can't quantify that though. I switched from Seagate after they seemed tosuddenly go down hilland i had heaps of failures. I've had similar amounts with Samsungs and am trying Hitachi now, so far so good. I'm mainly rtalking about 250-500gb drives as thats the bulk of what i use.

[wtbservices]

We have seen a slight uptick in bad drives but just a slight one. We use both Western Digital and Seagate.

A few years ago we were having an issue with a large number of drives failing within a few months of being installed. I called our distributor and asked about percentages of drives being returned hoping to find a more reliable brand. He looked up our return rate compared to the overall return rate and I was shocked to learn that we were returning 12% of our drives compared to the overall average of 4%!

The distributor got us in touch with the technical people at Seagate to try to help us figure out what was happening. Once we told them that the drives were failing after a few months of use and how they were failing they told us it sounded like the head was skipping. When the computer is bumped while the drive is in use the head skips across the platter like a rock skipping across pond. The drive can and will compensate for the very minor scratch on the surface but the microscopic particles it dislodges will float around inside the drive creating problems. Since the heads float over the platters at millionths of an inch even a slight bump might cause problems. Laptop drives have some mechanisms in place to protect against bumps but most desktop drives have none.

Our techs had this really bad habit of hooking up a system and getting it going doing a data transfer or a virus scan then sliding it out of the way to make room for another system. In many cases this was not a gentle procedure. We educated them on the matter and told them to never move a running desktop system. Thankfully, our return rates dropped back to the overall average with just this small change.

I don't know if this helps any but I thought I would throw it out there.

[neuroncs]

This is an old post but I gotta throw my two cents in here.

I know exactly what you guys mean about these hard drive failure rates. I experienced this first hand across all brands with the consumer-class drives - it doesn't matter.

The biggest difference I have found, is Consumer-class vs. Enterprise-class hard drives. When my TiVo drive died (crappy consumer-class AV drive), I replaced it with a Barracuda ES (Enterprise series) and it is rock solid for the past 2.5 years with zero artifacts in video playback and zero issues.

When my 8TB RAID-5 array went ballistic with Barracuda 7200.11 (if you'll remember - those are the consumer-class Seagate drives with the horrible firmware issues), I replaced the whole thing with WD RE3 1TB drives and they have been *rock solid* with absolutely *zero* issues (and i'm not gonna even knock on wood here). On a side note, never build RAID arrays with consumer class drives - they simply aren't designed for that and it's very hard to get away with it.

So in a nutshell, these are the rules that I abide by when it comes to hard drives:

1. Cooling, cooling, cooling
2. anti-vibration, anti-vibration, anti-vibration (think, rubber grommets!)
3. ENTERPRISE CLASS, PERIOD. Don't pay $40 for a 1TB drive that will fail in months. Pay $100 for a QUALITY drive that will fail after 5 years if properly taken care of

The cheap-ass parts they use in consumer class drives sickens me. They should all be pulled off the market and banned. It's very sad how clients buy off-the-shelf PCs and store dozens of gigabytes of family photos, documents, research papers, etc etc only to see it all magically disappear one day for absolutely no reason. I never build PCs with consumer-class drives.

[natrat]

Great post neuroncs. Have to agree especially with the RAID comments - built my 4 drive RAID5 with standard Samsng consumer drives and in 8 months three failed. Put Hitachi enterprise drives in and none have failed 6 months later. Have also started putting enterprise drives in all PCs - its a good selling point too.

[pyro77]

Western Digital are the only SATA drives we sell anymore. That does not include the Green or Blue drives(junk). Black for some but two months ago I instructed our order desk to only order RE3 or RE4 models which come with a 5 year warrranty! Smokin' fast as well!

I also am tired of hard drive failures. About 3 years ago, Maxtor won the Prize for ALL TIME with their drives which had about an 80% failure rate! Glad they are gone.

[nattivillin]

I don't even trust raid anymore. The technology is solid, but the error correcting of the OS makes raid unreliable. You can have 3 good drives, and 1 bad drive, and still have a system that will not boot / run properly. It just copies corrupt data to all the drives. Bah humbug!

[natrat]

I mostly agree re RAID. I've started moving to imaging software and internal/external backup drives for many small biz clients. Software like Shadowprotect can do incrementals every 15 minutes with very little overhead and also with excellent bare metal / hardware independent restore options. I've had to restore onto similar hardware for one client and it was straightforward. I do run RAID5 on my own PC though and its been great, long rebuild times when a drive fails though, the system runs like crap for over a day while its rebuilt onto the new drive. Next time I'd still probably ditch RAID and get a fast SSD for the OS. One of my contractors built himself a box with a SSD for the OS and it boots to the desktop in under 10 secs.

[natrat]

Heh - as i typed the above I've noticed a warning on my Intel RAID monitor tool - one of the three drives has had a SMART event and may fail shortly. That's the third drive i've had to replace in 6 months (2 x cheap Samsungs, 1 x cheap Hitachi).

[DynamicIT]

Hi Team,

To those people using RAID5 - do your research. It's a HORRIBLE system (less horrible with an expensive controller) that I happily used until a few years ago, not knowing any better. If you can fit 4+ disks in, use a RAID10. The disk cost is higher, but the speed cannot be compared.

Cheers
Mike

[bkendall]

I would second what Mike says. Have a look at this link for more information http://www.cyberciti.biz/tips/raid5-vs-raid-10-safety-performance.html

Brad Kendall
www.bradkendall.ca
brad@bradkendall.ca