What is the best hard drive brand?

[thebigcheese]

Too soon to tell. Eventually, you'll start to run up against the write count limits of the flash parts. Up until that point, it's anybody's guess how reliable they will be. I would expect them to be much more reliable because there are no moving parts. Again, though, there isn't enough data on them yet to draw any real conclusions, IMHO.

There's a limit on how many times you can write to the chip? That's interesting... But yeah, I was definitely operating on the assumption that only the moving parts bit would make the difference.

I have a Seagate in my desktop right now that I haven't used much (because I don't use the desktop much), so I might just buy an enclosure and an eSATA card because all the enclosures on Newegg seem to be 1394a only, but I think putting the hard drive on a different bus would help my recording performance anyway. Even after all that, I think I would still save a good $60 now and probably even more later.

 

[TyphoidHippo]

There's a limit on how many times you can write to the chip?

It's come a pretty long way in the last few years, though... I remember my first thumbdrive, lol. It held 16 MB and it lasted like 2 weeks. The 4 GB one I use now has lasted for almost a year already, I think (knock on wood).

I've got high hopes for SSD...

 

[dgatwood]

There's a limit on how many times you can write to the chip? That's interesting... But yeah, I was definitely operating on the assumption that only the moving parts bit would make the difference.

Yes, there's a limit to how many times you can write to a given cell. The flash controllers themselves use a technique called wear leveling to try to maximize the life of the part by remapping the logical blocks to different physical cells so that each cell gets roughly the same number of writes, but that still only goes so far.

It's come a pretty long way in the last few years, though... I remember my first thumbdrive, lol. It held 16 MB and it lasted like 2 weeks. The 4 GB one I use now has lasted for almost a year already, I think (knock on wood).

I've got high hopes for SSD...

Thumb drives are not the same as SSDs, and thus are not likely to have nearly the same level of failures. The #1 cause of thumb drive failure is failed solder joints at the USB connector, followed in short order by failed solder joints caused by board flex. Those failures are dominated by failures of what amounts to moving parts.

With SSDs, by contrast, they're mounted inside a machine, so they shouldn't have the same sorts of connector stress problems.

If you want to get a thumb drive that will last, I'd probably something where either A. the cap either swings up from the side or B. the USB connector slides into the body. Avoid the shove-on/snap-on cap designs like the plague. They just have too many problems, mainly because of poor cap design that causes the cap to add additional torquing stress on the USB connector.

 

[TyphoidHippo]

Thumb drives are not the same as SSDs, and thus are not likely to have nearly the same level of failures. The #1 cause of thumb drive failure is failed solder joints at the USB connector, followed in short order by failed solder joints caused by board flex. Those failures are dominated by failures of what amounts to moving parts.

With SSDs, by contrast, they're mounted inside a machine, so they shouldn't have the same sorts of connector stress problems.

If you want to get a thumb drive that will last, I'd probably something where either A. the cap either swings up from the side or B. the USB connector slides into the body. Avoid the shove-on/snap-on cap designs like the plague. They just have too many problems, mainly because of poor cap design that causes the cap to add additional torquing stress on the USB connector.

Good advice, thanks. I'm pretty sure that my original one, and another 64 MB one I had later progressively failed because I wrote to them too many times, I was able to (quite literally) watch the amount of space they had/reported in nautilus and/or thunar rapidly go down to nothing on two occasions that I can think of. That's why I think what I was saying was relevant. edit:SSDs and thumbdrives do have the same kind of limited-write compact flash memory, don't they?

I did have a few with lids that just plain broke, though, for the reasons you wrote, I guess. One of them was pretty solid metal, too...gave me a false sense of security The one I have now is one of the plastic verbatims that slides out. It's a tank.

 

[357mag]

I guess flash drives do have a limited number of write cycles. If solid state drives are going to have this limitation I can see a problem developing. Will there be a warning message presented when you have reached the maximum number of writes? Probably not.

More than likely you will realize the drive can't hold information reliably any more after it is already too late.

 

[dgatwood]

Good advice, thanks. I'm pretty sure that my original one, and another 64 MB one I had later progressively failed because I wrote to them too many times, I was able to (quite literally) watch the amount of space they had/reported in nautilus and/or thunar rapidly go down to nothing on two occasions that I can think of. That's why I think what I was saying was relevant. edit:SSDs and thumbdrives do have the same kind of limited-write compact flash memory, don't they?

I wouldn't expect the amount of space to decrease. I't expect you to just suddenly start seeing write errors. I suppose you might see a filesystem start marking those blocks as bad, but I've never seen that actually happen IRL except in the days of floppy drives on a Mac....

In Linux, seeing the amount of space decreasing, in my experience, is usually caused by a bug in ext2/ext3 that causes the dtime field to not get filled properly when an inode is marked deleted, thus causing the space to never get truly freed until you fsck -f the volume.

Either that or if you were using FAT... well, I never could get Linux's FAT implementation to work reliably. Admittedly, I haven't tried in a while, but it wouldn't surprise me if loss of space were due to their FAT support simply not being particularly robust.

But yes, SSDs and flash sticks do both use flash parts, usually NAND flash. To wear the flash parts out, you should have to beat on them pretty much continuously for several months even with early flash parts---several years with modern ones---so I suspect you were seeing something else at a higher level, but I could be wrong.

 

[dgatwood]

More than likely you will realize the drive can't hold information reliably any more after it is already too late.

You'll start getting write errors. The only data you would lose would be any files that happened to share the same cell with the data you were rewriting at the time. At that point, you'd know to save your work on another volume and start copying everything off the device because continuing to write to the device would be an exercise in futility.

It's not ideal, admittedly. Fortunately, it is likely measured in years of continuous thrashing with the better parts, months with the cheaper parts. For something like a thumb drive, you probably won't ever reach the limit unless you boot from it and use it for your paging file.

 

[TyphoidHippo]

I wouldn't expect the amount of space to decrease. I't expect you to just suddenly start seeing write errors. I suppose you might see a filesystem start marking those blocks as bad, but I've never seen that actually happen IRL except in the days of floppy drives on a Mac....

In Linux, seeing the amount of space decreasing, in my experience, is usually caused by a bug in ext2/ext3 that causes the dtime field to not get filled properly when an inode is marked deleted, thus causing the space to never get truly freed until you fsck -f the volume.

Either that or if you were using FAT... well, I never could get Linux's FAT implementation to work reliably. Admittedly, I haven't tried in a while, but it wouldn't surprise me if loss of space were due to their FAT support simply not being particularly robust.

But yes, SSDs and flash sticks do both use flash parts, usually NAND flash. To wear the flash parts out, you should have to beat on them pretty much continuously for several months even with early flash parts---several years with modern ones---so I suspect you were seeing something else at a higher level, but I could be wrong.

Yea, they were still factory-formatted FAT. It's worked great in Linux for a good while. I always leave them FAT so they will work on any machine I take them to (school, work, friend's house, etc.). The ones that shrunk down to nothing also appeared as near-zero capacity removable devices under Windows, and I did try to fix the filesystem every time it happened. edit: well, to be fair, the shrinking capacity-deaths only happened twice that I can think of. The very first drive, and a Sony something-or-other two drives ago. The other deaths were likely from actually breaking the way you explained earlier. The drive was just toasted. I remember reading many threads where people were discussing the same issue remember when it happened with the Sony one I had - they had been written to death and were no longer usable. I might look for those threads again later, just for conversation's sake. Either way, I'm not trying to say that I think flash memory isn't long-living these days. I was just giving an example of how I have experienced it become much more reliable over the years, to support why I said I have high hopes for the SSD becoming a good replacements for the HDD.

 

[RWhite]

I've been working as a PC tech for more than 20 years, and a PC salesman before that. While I have seen particular models of hard drives that had a high failure rate (one infamous Maxtor 20 gig comes to mind), I have never been able to draw a conclusion that one brand is significantly more reliable than another. Seagate, Maxtor, Western Digital, Fujitsu, Hitachi, Quantum, etc. I've probably owned more Western Digital drives than anything else but that has more to do with pricing and availability than any personal preference.

I have never heard of Lacie, G-tech, or G-drive. So far as I can tell they are not drive manufacturers, I'm guessing they are companies that make drive enclosures and then put someone elses drives inside them.

With a little common sense (including always having a backup copy), there is no reason a hard drive shouldn't give you years of service. However I recently had a hard drive that was only a month old fail on me, so it DOES happen.

 

[dgatwood]

Yea, they were still factory-formatted FAT. It's worked great in Linux for a good while. I always leave them FAT so they will work on any machine I take them to (school, work, friend's house, etc.). The ones that shrunk down to nothing also appeared as near-zero capacity removable devices under Windows, and I did try to fix the filesystem every time it happened. edit: well, to be fair, the shrinking capacity-deaths only happened twice that I can think of. The very first drive, and a Sony something-or-other two drives ago. The other deaths were likely from actually breaking the way you explained earlier. The drive was just toasted. I remember reading many threads where people were discussing the same issue remember when it happened with the Sony one I had - they had been written to death and were no longer usable. I might look for those threads again later, just for conversation's sake. Either way, I'm not trying to say that I think flash memory isn't long-living these days. I was just giving an example of how I have experienced it become much more reliable over the years, to support why I said I have high hopes for the SSD becoming a good replacements for the HDD.

It is certainly possible that Sony built a device whose wear leveling sucks. I wouldn't put it past them. Most of the Sony gear I've owned had pretty serious problems with its design, hence the reason I no longer buy their gear. At all. For my worst enemy. One apparently-not-quite-ATA-compliant TiVo, one "We'll sell you a replacement power switch for your camcorder for $400+", and one Hi-8 camcorder that went through three complete video head replacements while still under warranty was enough to make me swear off their products....

No, I'm not bitter....

 

[dgatwood]

I've been working as a PC tech for more than 20 years, and a PC salesman before that. While I have seen particular models of hard drives that had a high failure rate (one infamous Maxtor 20 gig comes to mind), I have never been able to draw a conclusion that one brand is significantly more reliable than another.

Yeah, in my experience, craptasticness comes in waves. IBM (now Hitachi) had a wave with their Deathstar line a few years ago where the drives failed reliably. Right now my experience is that Seagate is having a bit of a bad patch, though given the year delay before you figure out that there's a problem, they may already have corrected the issues that caused the high rate of crashes I've seen (both in laptop and desktop drives). Hard to say.

At least one model of drive a few years go (I forget which one---maybe one of the IBM drives) reliably failed very close to a certain number of hours of use (you could just about set the clock on your computer by it). It thus became a common practice to use those things for in desktop machines first, then wipe them before using them in RAID arrays so that they would have a significantly different number of service hours. That way, when one failed, you didn't reliably lose a second disk (and thus lose the contents of the array) while you were rebuilding the array with a replacement drive. (You would then perform a full backup, mirror it to a second array, and keep the staff on a 24x7 death watch to replace the remaining drives at regular intervals....)

I used to have a stack of dead Quantum drives that you could just about build a dollhouse out of. The FireBall series was a nightmare for me. Never had a single one last more than a year or so (over the course of several years). The number of hardware failures for Maxtor seemed to go way up for me after they acquired Quantum, and ditto for Seagate after they acquired Maxtor. Somehow, I've concluded that Quantum is like a corporate technology virus that infects everyone who comes into contact with their remains....

 

[iownrocknroll]

SEAGATE SAS, 15,000 RPM, RAID 5....Dell MD1000 Storeage vault with SAS connection to host.

 

[the-audio-man]

To be honest, some of the brands you list ie Maxtor actually has a seagate drive inside them. I owned an external Maxtor opened it up after it fell and broke and it was a segate drive. I wouldn't be suprised if a lot of companies are using one of maybe a handful of drives regardless of the external branding

 

[dgatwood]

SEAGATE SAS, 15,000 RPM, RAID 5....Dell MD1000 Storeage vault with SAS connection to host.

Don't get me started on Seagate these days. They're having firmware problems that cause drives to stop working (with data intact, but inaccessible). They posted a firmware updater to fix the problem. I'm told that the first version reliably crashed and didn't do anything. The newest version reportedly makes the data on previously working drives suddenly become inaccessible. It also tends to do it to every drive in a RAID set all at once, making recovery impossible.

Yeah, for at least the next year or so, I would suggest avoiding Seagate drives at all costs, along with any drive cases that contain them....

P.S. If that RAID setup contains recent drives, I'd back it up nightly if I were you.

 

[faxquo]

Hard Drives

I've had 3 Seagates crash on me .....

My vote goes to Western Digital (haven had any failures as of yet)

Also, I thought that Pro Tools doesn't support Raid configuration drives...am I wrong?

 

[TheGuitarMan]

My Western Digital RE3 drives are absolutely rock solid. I own three of them. Two in a Raid 1 array and one for backup purposes. Of course they're "Enterprise Grade" drives so they're more expensive, but over the course of a drive's lifetime $50 is nothing.

I've never had a WD drive go wonky on me. I've had a Maxtor start making the "Click of Death" noises then mysteriously start working fine again. I lost confidence in that drive immediately, of course, and plopped it in my parts bin.

I think a lot of people have premature drive failures simply due to careless shipping practices.