Backup drives
RAMWolff
Posts: 10,212
Hi,
I've been getting these little taps on my shoulder to look into a new backup drive. Mine are all mostly past the 3 year mark including my RAID so it's time to get with the program and not ignore what may be coming. No thanks. So I need valid input. I need a really good PHYSICAL backup drive that's dependable, long life and hopefully affordable. I used to be a HUGE Seagate fan but after a few issues with their drives I don't know... maybe they got the bugs worked out by now.. same with Western Digital.... too many issues but MAYBE they too have their "you know what" together after all these years. My main drive is a SS drive by Samsung and it's less than a year old so not too worried about that drive as of yet but yea, all my extra drive spaces like "D", "G" and "F" are all over 3 years old. I keep my Studio backed up there as well as content creation files like raw human photo files and projects in the works and copies of my completed ones. So for me a backup drive is essential.
Thanks for any help in making this important choice.
Richard
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
I don't think you have much choice really. From what I recall there are only 3 manufacturers of HDD's now: WD, Seagate and Toshiba. But if you're willing to pay significantly more $$ for an SSD, then you have some other choices (though I'm sure there are only a few SSD manufacturers too, who sell them under different brands?).
In my experience you have to have a backup to an HDD because they tend to fail. Doesn't matter the manufacturer, you need a backup. The HDD is really the only computer component I've had fail on any of my desktops over the years.
I just saw a $250, 1TB SSD that looks nice, but it's $250.
You pay your money and you take your chances....
Thanks so much. I kinda figured that but was hoping some more definitive experiences with life expectancy!
There really are only two choices now. Spend the money on toshiba/WD/seagate for a 5-year warranty drive and an external enclosure - and monitor the drives with a SMART monitor - or buy cheap USB drives that power down when not in use - and monitor them with a SMART utility. I've gone the cheap route - and have two pair of backup drives I alternate between weekly. FWIW, IBM is getting drives from all three above for their high-end storage subsystems. The main difference is that the subsystems all run raid with full time monitoring and automatic array rebuild, with a minimum of 8 spare drives and a call-home for replacement on the third failure.
I wish I could afford one for home use.
I prefer a 'smart copy' utility over a 'backup' utility because it is much easier to verify the result (an unverified backup is a non-existent backup).
You can find some stats on how long drives last if you're interested...something like "80% last for 4 years" is one I saw. I also saw one stat that WD is much more reliable than Seagate, but the failure I had this year was a WD.
You pay your money....
I think you're going to get a great deal of "folklore" in your responses... "SEAGATE is the only brand to buy. Every WD drive I've *ever* purchased has DIED on me!"... "NO! SEAGATE SUCKS!! WD drives are inferior for <insert barely relevant techno-jargon here> reasons!"
I've got a Seagate and a WD on my desk, and they're both rolling along. I've had backup drives from both fail. Early failure probably has more to do with particulars of one manufacturing run vs. another than anything else.
If you want more piece of mind, compare reviews on Amazon and New Egg (or any other source that sells a large number of units.)
(The only time, btw, that you'll benefit from having an SSD back-up drive is if both sides of the copy chain are fast, otherwise you're limited to the maximum sustained/intermittant read from the source or the maximum write sustained/intermittant on the BU drive.)
Yup. Seems like with any kind of tech stuff there's myth and folklore that just won't go away. And I think the latest one is that SSD's have a really short lifespan, measured in minutes or something. I suppose time will tell, but that's another myth I think is baseless.
So IMO, if you can afford a 1TB main SSD drive that's the way to go, with maybe a HDD backup (or maybe a portable USB).
OK, after doing some research the HGST drives seems to be the most stable.
Here is a link to some info on WD, SG and HGST and Toshiba drives. The info and graph's tells the story for me!
https://www.backblaze.com/blog/hard-drive-reliability-stats-q1-2016/
Scroll down to the Hard Drive Failures Rates by Manufacturer graph and see what I see....l
You can already write off that myth about SSDs too. They're being put into major enterprise hardware now as the primary cache, and even primary storage in some units. Granted these are the ones with massive redunancy and 24/7 monitoring built into them, but at the prices this stuff is the manufacturers wouldn't use them if they were likely to fail early. My office is about to get a new storage array that is all SSD.
On a point more on topic, I've been professionally on the technical side of the alley for about 15 years now, and a hobbyist for about25. In my experience, there really isn't a particular major brand that is enough more reliable over the others to worry about. Yeah, I probably wouldn't go with some no-name brand, but the big 3 are pretty much on par from what I've seen. If you held a gun to my head and made me make a choice, I'd probably go with Toshiba, but I also haven't had as many of theirs as I have the other brands.
Sorry, but IMO those statistics are pretty much irrelevant for the average user. First, you're talking about data center drives that are likely under completely different usage from average users. Second, you're talking about failure rate per manufacturer...that doesn't tell you how long a drive you buy from one of the three manufacturers will last. Only that, out of tens of thousands of hard drives over however many years, 5-10% of them failed. It doesn't talk about the particular model, the usage, how many years, and so on.
Seagate had FAR more drives in service (15,000 to 36,000) compared to WD (600 to 1,700) compared to Toshiba (60 to 240). It's no wonder that Seagate had a higher failure rate since they were far more likely to have a broader representation of all types of drives and failure rates.
It's a bit like saying "I had a brand xxx hard drive that lasted for 10 years, therefore brand xxx is the best"
You pay your money and you take your chances
The Toshiba and WD stats are useless, but HGST vs Seagate is pretty telling.
I think the Seagate may be the only statistic that is even close to realistic, only because it's a large and representative sample size. You can't compare the rate for 25,000 drives to the rate for 60 drives. Statistically, 60 drives is not at all representative of a manufacturers' entire production. A statistician would probably laugh at that comparison. I'm guessing that if you compare 25,000 drives of all three brands the failure rates would probably be similar.
??????????????? They had tens of thousands of drives for HGST.
Yeah, but HGST is Western Digital, and it's their enterprise drives. I suppose if you're going to buy an enterprise hard drive (I assume they cost big bucks), then yeah, maybe it means something. Otherwise, not so much.
So when the failure rates are shown to be different, tens of thousands of drives aren't enough anymore?
No, they're not big bucks.
Yes, HGST is a subsidiary of WD. Since there is no good WD-in-general data, we cannot say if WD drives have an equally large advantage over Seagate. We can say HGST does.
Wow, you're right. You can buy a 1TB HGST at Newegg for $40. I wonder why they have a failure rate that's so much lower. Looks like the Ultrastar are their cheap versions and Deskstar are the more expensive ones. I wonder if the ones cited in the report are the high end versions or something. Hard to believe that WD would have two different qualities of hard drive technologies and sell the better ones at a low price.
Since they were a buyout, not something spun off of WD, it's likely/possible they still have completely different manufacturing and even development.
EDIT: Looks like Megascales and Deskstars were tested from the product numbers.
In the comments down below the article they said the company that did the article said there wasn't much difference between the enterprise and consumer brands of the HGST drives.
Damn, I just bought a 3TB drive. Wish I had known about HGST.
Although in the scheme of things, chances are that this drive will last until I upgrade this desktop in 3-5 years, and since they're so cheap it's not that big a deal.
Depends. Are they both rated for the same number of i/o operations per second? The same write/read ratio? The same idle/access time ratio? There are a lot of things that can be played with to change costs.
As for failure rates by manufacturer - there isn't any guarantee you won't be that .01% who gets nailed. Remember, you're up against Murphy. And the ONLY way to win against him is multiple redundant backups. The last 7 years of my professional career were backup/recovery and disk provisioning guru.Been there. Scars have mostly healed. :-)
So namffuak had you heard about HGST, since you're in the business? Do they have a reputation for high reliability?
They're after my time, apparently. FWIW my main system has two 2 TB and 2 1TB internal seagate, a WD 500 GB, and a 500 GB ssd. And 4 external usb drives - two 4 TB WD and two 2 TB Toshiba. My other home system has two internal SSD drives - 128 GB and 256 GB (a refurb Dell 7010, came with the 128 and the 256 was pulled from my previous laptop) with two Tosiba 1 TB usb drives and two 4 TB WD usb drives for backups from a third system and redundant copies off my laptop.
The last 3 drive failures I've had were a 1 TB internal seagate that developed 10 MB of nonwriteable sectors while under warranty, a Maxtor (forget the size) that died the death a month after the warranty expired (I'm 3 for 3 on Maxtors doing this to me over my lifetime) and a 4 TB WD MyBook external that the SMART utility flagged as imminent failure.
As far as the office environment was concerned - I didn't do desktops, just the big iron. And even with that, there were days I'd show up and see an email from the disk system about a drive failure, an email from IBM Tucson about a failure and (if required) dispatch of new drive, and then get a call from my CE saying he'd be in around 11:00 AM to install the drive(s). The first 3 generations of the ESS would get the drive replaced upon failure; the fourth generation, IBM shifted to replacing when the third drive failed (default config was 8 spares). And the nice part of the ESS was that the new drives went in as new floating spares - no raid rebuild or copy process.
Those IBM disk arrays are awesome beasts. A few years ago my office took a nasty power spike when an idiot vendor screwed up replacing the batteries in our server room UPS. That spike fried 3 production servers, several decomissioned ones that were still live, and 13 drives in our ibm XIV array. The array just tossed out emails to everyone it needed to, reshuffled the bits around on the surviving drives, and kept right on trucking until IBM got new drives in for it. We did not lose a single bit out of that array, even with that many simultaneous drive failures.
And the DSS 8100 is even more so. And priced accordingly. :-) And in my shop I had two of them - with a couple of key datasets software mirrored between the two. Shortly after I retired they moved to the XIV - the 8100s were massive overkill for us (and for the rest of you reading this - it's pronounced ex-eye-vee, not 14
) Our busiest array in the busiest 8100 was running about 22% read and 7% write - I understand the XIV has been running closer to 65% read (peak) and 25% write (also peak) - so lots of available bandwidth stil.
Priced accordingly is one way to put it, we had to move off the XIV to a hitachi system when our maintenance contract on the XIV ran out because it was way too expensive to replace. We're getting ready to move off the Hitachi now onto some new system that's essentially a server room in a box. It's amazing how far enterprise tech has come. 10 years ago we had 20 racks full of gear that we were totally thrashing, we're about to be down to 2 racks (and that includes our tape library) with plenty of overhead in the system, while doing a hell of a lot more business than we were 10 years ago.
From the data it seems that HGST drives are a offshoot of WD and seem to be their Enterprise drives from what I've been able to find out. On Amazon they all have some pretty good reviews. I found a 3 terrabyte (which is what I'm after) for like 60 bucks with a 65 meg cashe, which is pretty nice.
For ordinary consumers I think it is important to differentiate between backup and archives. Sometimes I get the impression that many are using the terms synonymously. A backup is naturally just that: a copy of important data set aside in the event of drive failure on the device that holds the "working copy" of that data, allowing the data to be restored to a replacement drive from the backup. Archived data tends to be different: archived data is data which is considered to be important enough for some reason to be retained, but for which there may no longer be an immediate need or use. Storing such archived data on a "backup" drive to clean up or make more room available on a drive on another device makes sense, but doing so does not constitute a backup, since the archived copy represents the only copy under such circumstances. A second copy of this data is required somewhere else in order to hedge against the failure of a backup drive. I only mention this since RAMWolff has expressed concerns over potential failure of his backup drives. Such failures should not be catastrophic if they only contain backed up data. The latter can always be restored to a replacement drive or drives from the original source. But archived data, if it only exists as a single copy, would naturally be lost if a backup drive failed. So the point is, backup drives should be dedicated to that purpose - basically spare copies of important files. You need at least two copies of everything that is important to you, somewhere, since drives will all eventually fail and no drive manufacturer is going to guarantee the integrity of your data.
I've got 5 Seagate External Hard Drives (all desktop models) and 0 have failed...knock on wood. (Note: The last one is a 6T Backup Plus HUB that I've only had for a few weeks so 4 of my drives have age on them.) This newest one is my favorite so far. I paid $119 for a Seagate 6T Backup Plus Hub ($80 off regular price...sale still going on) at Amazon. I'm pretty impressed so far as it's quiet; has extra 3.0 USB slots; runs cool; lots of storage; plays well with Windows 10. (This one can also be used with Mac.) It has excellent customer reviews providing you have a 3.0 USB to plug it into. Reviewers say that there is an issue if you plug it into a 2.0 USB. (I will say one of my older Seagate's runs hot...reason for mentioning that this newest one runs cool.)
Whatever, you end up I'd recommend looking for something that has Hub...so you get those extra USB slots. I'd also look for something with 3.0 USB...quite a few I looked at only had 2.0 USB.
Given the choice, my next HDD would probably be an HGST. But considering my usage is likely vastly different from a 24/7 enterprise usage like in the article, I doubt that the failure rates really apply to me. It would probably take 20 years of use for my usage rates to come even close to what an enterprise drive does in 1 year. And since I always provide at least 1 backup, HDD reliability is pretty much irrelevant. It's not like I won't buy a backup drive because I'm so convinced my HGST drive won't need a backup.
I've used WD disks only for the last 15 years, bought over 50 during that time I think. I think about 12-15% have failed during that time though some first after several years. A couple have had bad sectors but after they were reallocated they have been running fine since.
The worst problem I've seen with their newer drives is weak sectors - sectors that are readable but are very slow to read (many retries). One was simply unusable because of a lot of weak sectors in the area where the MFT was located, which made it extremely slow. I couldn't figure out what was wrong as WD's own diagnostic tools reported it was OK. I changed cables, SATA ports etc. but nothing helped. Then I came across HDSentinel and ran a full surface test, which revealed all those weak sectors. I've attached the test results - all the dark blocks have weak sectors, the darker the weaker they are. The performance test shows a similar poor performance in those areas. A disk like that should not be able pass QA IMO.
I actually put a hub in my cart and then saved it for later. I have a couple of enclosuresures that I like allot and I plan on putting the HGST 3 T into that after I move all the content over from it to the new drive as it's my main backup. I'm thinking of getting two at that price actually. The other internal HD I have is aging and that's my MAIN DAZ Studio drive so that's got me a little worried as well.
Hmmm, thats VERY interesting to me. I DO have a HD that does check out fine but when I access it it takes a while to show my files but I don't hear any grinding or any extra noise like that from it so I thought "I have no idea why this drive is acting this way" sort of thinking. Is there a link to the tool you mentioned? Also does it repair these areas or is this just an informational tool?