Apparently a today-only deal at Amazon. I just bought two of them.
http://www.9to5mac.com/node/5102
http://www.amazon.com/dp/B001NGOIJO

<CLICK>

Thanks!

Damn, and I just bought four of these for $115 each.

Make sure to update the firmware on these in case they are the older ones that may become inaccessible one day. link.

Just canceled my order after reading this.

Yikes! I'll pass on that.

First I've heard of that problem. I've got about 8 ES drives and I've not had any problems with them. I had to update the drive firmware to prevent the startup problem from occurring but thats it.

/me would like more specificity about that particular clause.

Yup. This is from a company that recovers blown hard drives, so they may not entirely know what causes the problem, but I'm impressed that they're willing to talk so candidly about what the problems are and how they go about fixing them. They don't talk quite so candidly about their recovery prices.

Its a drive recovery firm as well. They're going to make out that it is a dire situation and that they're the only people around who can recover your data.

Ehh. Its advertising. Here is a list of things we can fix. The techniques involved are fairly standard.

My 1.5TB drives from the first available batch are running fine, though I've been meaning to update the firmware for a while. Of course, they're in the ReadyNas and have a spinup count of 6 or something last time I checked.

The interesting thing with these drives is that the size is so large, their started error rate makes it highly likely that you'll hit unrecoverable error while copying the contents off once.

I'm glad I'm not a hard drive designer, it's really amazing that they work at all.

Mmm.. the infamous bug-of-death Seagate 1TB drives.
I wouldn't touch those with a ten foot pole, really.

WD seems to be the only safe brand for large capacity at the moment.

Perhaps in a year or so, all of the b0rked Seagates might be gone from retail channels, but until then it's a total 1/256 crapshoot at every reboot.

-ml

Note that the bug is even more widespread than Seagate admits. I have seen (in person) drives fail in exactly the same way as the acknowledged ones, but yet Seagate claims those to be "unaffected".

Yeah, right.

Wow, that sucks. I'm glad I have them in a mirrored array...

mlord, can you expound on why it's 1/256 crapshoot when booting?

The various vague descriptions of the bug all point to an 8-bit event counter in the firmware. Whenever the events it has been counting hit the wraparound point, it becomes toast and never boots again.

But this only happens if the system is rebooted or shutdown (one, the other, or both.. I forget), at an instant when that particular event counter happens to have the magic number in it.

The events that trigger the counter to be incremented sound like they are slow events .. like daily or something. But after a while, it becomes a bit of a crap shoot as to whether your next reboot/shutdown coincides with the bad counter value.

Good luck.

The explanation I heard was that the drive keeps an internal log of events like remapping sectors. If the log hits 320 entries when the drive restarts, the firmware doesn't like it and prevents it from starting properly. Seagate have some sort of internal diagnostic tool or program that can reset the log and then allow them to reflash it to a working firmware.

Hi,

Has anyone seen this on the 750 GB Seagates?

I have 4 of them in my system. I had a set of PATA version 750s for over a year and they were fine although, I do not power down my system unless I need to. I replaced the set of PATA drives with SATA versions and they have been running fine for months.

Any data on the 750s being affected?

Thanks,

Ross

I hope that's not a matched pair. I've recently taken to buying a selection of drives, so that I never get two from the same batch.

My server will shortly have 1TB drives from Seagate, Hitachi, WD and Samsung in it.

I recently bought a 1Tb drive and gave Seagate a miss. The pricing was very similar (AU$140) so WD it was!

Yes. A buddy of mine here upgraded all of his drives in Nov-Dec 2008, with several 750GB, 1TB, and a couple of 1.5TB drives.

Three of them (one of each type) suffered the firmware breakdown (dead drive) bug, and got RMA'd for replacements. And for good measure, he sent the others back as well, after receiving the replacements for the first three.

Ugh. Avoid Seagate everything like the plague, for at least a year folks. They're behaving just like they did in the stiction days, because that's how to maximize profit in the face of a major would-be recall.

-ml

Yes, but that's a firmware bug with a known fix (linked earlier in this thread), right? If one applies that firmware fix, the bug is gone, right?

Yes. Seagate did screw up the initial update though for a few drives but they were fixed with another update...

One problem is that there are more drive versions affected than those they officially list on their site. You just have to wait for those to die suddenly before cursing the purchase decision.

Unreliable junk.

Cheers

I went to that link, and it appears as though the $79.00 drives are not the duff drives that require this firmware upgrade. DriveDetect.EXE tells me that the drive I just bought is an ST340810ACE.

Does anyone have any knowledge if that model number is one of the affected ones?

Also, the manufacturing date code on the drive is from February of 2009, if that's a factor.

Probably okay. I guess you'll just find out some morning. Or not.

-ml

Well, now Amazon is carrying the 1TB WD drive ("EcoGreen") for $93.80, versus the aforementioned Seagate, now for $88.

Ahh.. avoid those "Green" drives as well. Or go and download the *MS-DOS* (WTF?) utility from WD and fix them with it after purchase.

They ship with a "frequent head unloading" feature enabled, which will kill a drive when continuously powered on after about a year or so.

Cheers

Can you provide some documentation on that?

Mark is our definitive IDE expert! You dare question him?

I've stuck a couple of 1TB Samsung EcoGreen drives in my new ReadyNAS NV+ and haven't found anything on the web to say there's anything wrong with them long term.

We all remember from the Karma days how unreliable user-reported disk problem anecdotes can be. The head unload issue is reported by smartctl and so is almost certainly real: http://www.silentpcreview.com/forums/viewtopic.php?t=51401 although, equally, the variability of the anecdotes on that thread makes it sound as if some people don't get affected; perhaps it's improved in later firmware versions.

Peter

Really recent Linux distro's already "quirk" those drives by disabling APM (Advanced Power Management) on them using hdparm.

The best solution apparently is the wdidle3 MS-DOS utility from WD, with which the "Timer can be set from 100 ms to 25.5 seconds, in 100ms increments", as well as disabled completely.

-ml

Heh. Not questioning; I just wanted more data.

Modest confusion on my part: "EcoGreen" is the Samsung term. "Green Power" is the Western Digital term. The Amazon deal I posted was for the Western Digital drive.

If you follow the links from Peter, you'll find that most of the discussion is of the WD10EACS and earlier versions. Amazon is selling the WD10EADS. ("D", not "C", the key difference being a doubling of the cache size.) If you poke deeper into the SilentPCReview thread, the consensus seems to be that, on these later drives, the utility that lets you adjust the idle timer is no longer supported and using it will void your warranty. Also, those people are convinced that this drive is still parking the head every eight seconds but the SMART counter isn't incrementing.

Very disturbing.

So, dearest mlord, I need to buy some 1TB drives to hold my music and pictures. What brand/model do you recommend?

Heh.. no good choices, are there!

Definitely NOT Seagate. Doesn't WD have a non GreenPower version?

The Hitachi drives are also big, and very quick (noticeably faster than the other two). But I did have one of those die prematurely here last fall..

Gotta love industry consolidation.

-ml

If you go to NewEgg and ask for a 1TB internal SATA drive, you get ten choices:

- four WD models, priced from $99 to $209.
- four Seagate models, priced from $89 to $159.
- one Hitachi, $87.
- one Samsung, $99.

If you ask for 1.5TB, you get two choices: a Seagate and a WD. If you ask for 2TB, you get one choice: a WD. That's it. Here's the full list (assuming the link works).

Given that list, what's the least likely to go kaboom?

What about Samsung ?

I'm happy with my 2 1Tb samsungs so far. Very fast, pretty quiet, no deaths yet.

Unfortunately personal anecdotes over the reliability of drives aren't really useful, they are just a single data point. I was more after any known documented issues that Mark might have come across.

data points?

I bought 9 drives and had ~17 failures with the HD103UJ 1Tb Samsungs over a few months.

Eventually Samsung replaced all 9 with Enterprise edition ones. They fail far less regularly - only ~5 so far...

If I had to purchase my own 3.5" drives right now, I'd look for WD drives without the "GreenPower" thingie.

For notebooks, WD is also my top choice for a mechanical drive.

But I'm expecting delivery later this week of an OCZ Vertex for my main notebook. Hard to beat a good SSD when huge capacity isn't needed.

Cheers

Okay, then. Amazon is selling the WD "Caviar Black" 1TB disks for $105.98. I've ordered two of those. I don't actually care about speed so much as noise and power consumption (and not gratuitously blowing up, of course). From a pretty detailed review, here are some noise numbers:



The Caviar Black that I'm getting is 3.6dB louder than the Caviar Green. That's non-trivial, but tolerable (we'll have to see how good a job the case does of muffling it). But then there's power consumption:



Since most of the time the computer isn't actually doing much with the disks, but it's otherwise on, the idle power numbers are the ones that really matter to me. The power differences are quite surprising: an extra 2.7 watts for what's presumably very similar hardware. None of this is in "break the bank" territory. It's just unfortunate that I have to take a noise and power penalty just to have a drive that might actually last a few years without going belly-up on me.

My server's in the spare room, so the noise penalty means more to me. I figure that I'll be upgrading the disk well before the few years is up (I double my storage capacity about every 2-3 years). With regular backups (which, admittedly, I'm still struggling a bit with), I'm not that fussed about dead disks. If I was worried about uptime and high availability, my priorities would be different, but going without my server for a couple of days is do-able.

Weird, I guess the Velociraptor consumes so little energy because it's a 2.5" drive?

I have the 300GB version, and even though the velociraptor places highest on that noise chart, it hasn't bothered me.

By the way, the 1TB Green Power WD drive is on sale today at Newegg for $89

Of course, when we're talking about disk noise, we had an old Fujitsu Eagle (470MB !) in the CS undergrad association office when I was a student. That thing was unbearably loud, and Wikipedia claims it drew 600 watts.

Sometimes it helps to put things in perspective.

My master plan to solve this issue:

I presently have my data stored on a LaCie 2x750GB Firewire external drive. I have a second one that I use as a Time Machine backup.

The new hard drives, striped together and running inside my new Mac Pro, will be my primary storage. One of the Firewire boxes will be the Time Machine backup. The other Firewire box will be moving to my office and connected to my work machine, and I'll set up rsync, giving me offsite, networked backups. I figure that, by doing the initial dump while it's still at home, the incrementals shouldn't completely destroy my DSL line.

After Apple releases OS X 10.6, I hope to rearrange the Mac's internal drives to use ZFS. At that point, I can retire the Time Machine drive.

Interestingly, I'm still actually actively using a 202MB WD AC2200F hard drive that was the first hard drive I ever purchased, back in about 1992-3. It's obviously not got much capacity, and it's not speedy, but it's never thrown any errors and it still works just fine.

Considering it is only a $5 discount, they've got a huuuuge advert for that drive

Yeah, that's pretty much how all the bacon I get from them in my inbox looks like. I'd unsubscribe, but occasionally there really is a great deal, or I'll see a deal for someone I know. It's quite helpful.

These are now down to $99.99 with free shipping. How did these work out for you?

My computer locked up during backups last night because of lack of disk space, so I guess it is time to upgrade. I am far from a hardware guru, so it is with some trepidation that I consider this.

I think the two drives that I want to replace are SATA drives, but I don't know how to tell for sure. Device Manager says they are: "MAXTOR S TM3320620AS SCSI Disk Device". I know for a certainty that they are not SCSI drives, but how can I tell if they are SATA?

Assuming they are SATA drives, is it then just a matter of removing the old drives, plugging in the data and power cables to the new drives and copying the data from my backups? (The replacement drives will be for my D: data storage drive and E: backup drive. My C: system drive is a separate small IDE drive.)

In the meantime I have gone through and cleaned up my existing drives so that hopefully I can re-run my backups tonight and not have them crash for lack of space.

tanstaafl.

Just plunk STM3302620AS into google:

Maxtor DiamondMax 21 320GB SATA2 7200RPM 16MB NCQ Hard Drive

SATA controllers can/do appear in Windows as SCSI controllers for some reason. If you're certain that its not SCSI then it'll be SATA. Mark has confirmed it anyway.


Pretty much.

Quite well. These turn out to be precisely the disks that Apple sells you when you order a new Mac Pro with a 1TB internal drive. So far, they're working without complaint.

Dragging up an oldish post.

Is WD still the disk of choice for 3.5" drives? What's so bad about the WD GreenPower drives?

Stig

The greenpower drives spin up, spin down, spin up, spin down, spin up, spin down ... perhaps a hundred times a day.

Which means they'll hit their design lifespan within a couple of years or less.

That's not good for data safety.

Thanks, Mark. Is there a way to disable that feature? As the GreenPower drives are cheaper than Caviar Black. Or is Caviar Black a better drive overall? I plan to use them as drives in a NAS.

Stig

Do they spin down even if the host isn't telling it to do so?

Apparently they did in early firmware versions. I've just checked mine, though, and it's had 270 spin-up cycles in 2500 power-on hours, which sounds about right (two or three a day) for the ones initiated by the OS ("hdparm -S 242 /dev/sda").

Peter

I did wonder, I have Samsung EcoGreens in my ReadyNAS and looking at the SMART data and it only shows 7 power cycles and over 3000 operating hours which makes me think they've been spinning pretty much constantly since I bought them.

Power cycles aren't the spin up/down cycles though.

smartctl shows a Start_Stop_Count of 7 which is identical to the Power_Cycle_Count.

Check the head load/unload counts in the output from smartctl -a

EDIT: mmm.. that's probably the same as the "Start_Stop_Count", which Andy has already checked. Dunno for sure, though, as I won't allow any of those drives into our gear here.

"Load_Cycle_Count" is the one to look at.

On my eight WD10EADS-00L5B1 drives, the Start_Stop_Count and Load_Cycle_Count (raw) values are the same on the same drive, either 58 or 59. And they've been up and running for a little over two months.

Srsly. Load_Cycle_Count and Start_Stop_Count are both about 270, much less than Power_On_Hours at about 2500, and much more than Power_Cycle_Count at 17.

Peter

Okay, so that's an indication that they may have fixed the issue with newer firmware, as you suggested.

Good!

Either that, or they're just not reporting it any more in the S.M.A.R.T. data.

I have a WD10EADS and it shows power on hours 1380, power cycle 172 and start/stop and load cycle are both 174.

Coworker just had an ST3750330NS go brick on him. Suddenly. Right after a reboot.

Anyone know if that drive is one of the affected ones?

Newegg reviews don't look so hot.

... And if his is one of the models with the firmware bug, is there a chance he can recover once it's bricked?

You have to send it to Seagate for them to do their magic on it but it should come back with full access to data. I assume they have some sort of secret manufacturer programming interface which allows them to reset the log.

Thanks.

Wow, Seagate have been the Sony of hard drives this past year. Damn. I pretty much exclusively buy Seagate, so I'm knocking some wood right now.

I applied updated firmware to my 1.5TB drives as soon as I got them, which was a little while after Seagate finally released their fix. At the time I had to live-chat to get the firmware files as they weren't available for free download.

The HD has buggy badly written spyware that doesn't even work properly which is installed without your approval and it leaves a backdoor open the size of an oil tanker when you try to uninstall it?

I was referring to the Sony battery fiasco which was a much bigger deal than their root kit issue. Week after week and month after month we kept hearing about a new batch of batteries. So too with Seagate's drives. FIrst 1.5TB, then 1TB, now ES models at 750GB.

Since the rootkit fiasco, Sony has not received a dime of my money, and it is unlikely that they ever will again.

tanstaafl.

I've crossed LG off the list.

BlueBirds

Reviews

Re: Sony (new today)... See what I mean?

Of course, bad power adaptors aren't just a Sony problem.

No. Not really. If it was only Sony that ever recalled a faulty part that was manufactured by somebody else then yes there would be a point to you mentioning that.

I've got a bunch of replacement Dell laptop PSUs because the originals had a similar problem to these VAIO ones. My original Xbox has a special replacement mains lead that has a RCD built into it because the socket is flakey. There are a multitude of recalls for car parts which are even more dangerous because there is a serious chance of injury or death.

Whilst it is bad that they sold these potentially dangerous PSUs, it isn't specific to Sony. They're not even made by Sony. They asked one of the big PSU manufacturers that they'd like a PSU that outputs this voltage at this current and they'd like their logo stamped on the outside. I wouldn't be surprised if Apple used the same manufacturers.

Which naturally leads to the next question -- what's a decent 1TB or larger drive that isn't going to go south mysteriously?

If you scroll backward in this very thread, you'll see Mark Lord advocating and me buying Western Digital Caviar Black 1TB hard drives. They were also standard equipment on my recent Mac Pro.

I also bought a pair of the WD Caviar Black drives and have been completely satisfied. Of course, they've only been on-line for a couple of months, and one of them is only powered up for a few hours every few days or so for backups.

BTW, be wary of buying this drive from Amazon. Many of the reviews state that the drive was extremely inadequately packaged, rattling around loose in an over-sized box and arriving DOA. By contrast, Newegg shipped my drives wrapped in bubble wrap in a properly sized box filled with styrofoam packing peanuts. You'll pay about $15 more at Newegg, but it is money well spent.

tanstaafl.

tanstaafl.

Hi,

I have used 8 of the WD 1TB drives (4 each for system storage, 4 for disk to disk back-ups every 2 weeks) in my i7 SPICE simulation system.

They seem to work fine for me.

Heeded Mark's warning NOT to buy Seagate drives.

Ross

Cool... wasn't sure if there was any more recent recommendations, or if the newer Seagates were okay.

They don't seem to be. Even Patrick is having issues with a Seagate drive here now, and their firmware-of-death saga is less than a year old yet. Avoid.