Hacker News Reader

Comment by hn_submit | Hacker News Reader

This seems a little over the top to me.

Just use a magnetic hard disk and your data will be fine for many decades if not centuries. You don't need M-Blu Rays (which have very limited capacity anyway). A hard disk is cheap and stores several Terabytes.

You obviously need to store the data in a physically remote place like your shed, car or safe deposit box in case your house burns down.

Never, ever use a SSD if you want your data to be safe for any amount of time. What's worse, SSDs can fail unexpectedly with almost zero chance of data recovery.

View on HN

> Just use a magnetic hard disk and your data will be fine for many decades if not centuries.

A few months ago I needed to pull from an offline backup. The disk, which used CMR and was purchased new from an actual computer supply company (i.e. not Amazon), was burnt in for about 72 hours before it was loaded up with data and shoved in a padded box. Ten years later, the drive was pulled, and there were so many media surface errors that most of the data was unrecoverable. (I haven’t bothered opening the drive to see if there is some obvious sign of what exactly failed.)

In my case the backups were created in multiples using two different brands for exactly this reason, and the second copy was fine. But I wouldn’t say it’s the case a single magnetic disk can take the place of BD.

Mechanical disks are also getting expensive. The disk purchased to replace the failed backup was being sold at three times its original release MSRP. Some of this is due to the AI hysteria but some of it is also that the market is shrinking and becoming K-shaped.

Oh for sure don't trust any single copy. But I wonder, how many surface errors? How much of the data would have been lost if all your drives degraded that much?

For example, if most files are multiple gigabytes, the drive has 3 million tracks, and 3 thousand of those tracks are ruined: one degraded copy leads to most files being broken, two degraded copies leads to only a single digit number of files being broken, and three degraded copies means >99% odds of recovering everything.

And with the magic of par2 files, a single drive with a couple percent of corrupted tracks can be fully recoverable. Any cold backup I do gets some of those.

Ten years in general is way too long to leave any piece of electronics unpowered, let alone a fine mechanics device like a HDD with moving parts. They need regular exercise.

You should always let electronics acclimatize before powering them up. I'd say at least 24 hours.

Just use a magnetic hard disk and your data will be fine for many decades if not centuries.

Comically untrue, and trivial to verify that it is.

You obviously need to store the data in a physically remote place like your shed, car

Let's make it worse by keeping them someplace with terrible environmentals.

If you're gonna be wrong, don't be just a little wrong...

Tell that to my circa-2000s WD MyBook that clicks and fails to initialize after ~17 years in climate-controlled storage in its original packaging. I'm about to attempt surgery on the heads to see if I can get it going again.

I also have a binder full of DVD-Rs, of about the same age and containing mostly the same data, that all read perfectly with no issues.

You're asking for trouble if you count on just one backup solution to keep your data safe.

You have the option of recovering the data through specialized shops that perform surgery on the drive.

> Just use a magnetic hard disk and your data will be fine for many decades if not centuries

It is absolutely not safe for many decades. If you haven't encountered any drive failures yet — you're lucky.

> A hard disk is cheap and stores several Terabytes

You should check prices again, I wouldn't say cheap considering price bump during the last year (or around so). M-Disks are way less convenient, but pretty reasonable for their cost.

I've had one drive for almost 20 years (don't know how long exactly). It is powered up once in a while and stored inside the house. No read errors of any kind as far as I can tell.

Even if perfectly held still in a sealed box with perfect conditions, hard disk drives need controllers; they need the chipset to drive them, which has ROM and other stored microcode that is susceptible to bit flips due to cosmic ray impingement. I've traced faults in a few 20+ year old HDD's due to failed controllers likely due to such loss (either in the main controller fail, a ROM/BIOS chip fail, or less likely but still possible the cache IC, or motor controller). The data is there on the spinning rust, you just won't have a way to access it.

The failure of SSD's you warn of and the way the HDD's fail due to a dead controller aren't that different.

This won't happen with a physically written media like a CD, DVD, or Blu Ray disc. With physical media, you're separating out the media from the reader (the drive). If your disc reader dies or you cant find it, etc, you can go find any another one on ebay even decades from now, very likely.

Doing surgery on an HDD controller PCB to replace a dead controller, while possible, sucks, and may be incredibly difficult if not impossible to source parts for in the future.

Offline HDDs are not a reliable long term data storage. Not even for a few years.

Hard drives are a safe option if you keep them redundant, online, and scrub them regularly. They are not a good long-term offline archive format.

Hard disks aren't even rated for one decade.

Just picking one at random: The Seagate Exos X20 series of hard disk drives is rated[1] with a MTBF of 2,500,000 hours. That's 285 years.

I don't expect anything with moving parts to last 1/10th that long, but MTBF is MTBF. :)

[1]: https://www.seagate.com/files/www-content/datasheets/pdfs/ex...

MTBF is for hardware failure, not data loss.

This recent study I found [1] shows almost 4% HDD loss after only 2.5 years for consumer hardware, enterprise products being a bit better.

[1] https://research.cs.wisc.edu/adsl/Publications/latent-sigmet...

If the data storage hardware cannot read a block that it has stored within itself, then has the hardware not failed?

---

Suppose I hire a scribe named Gunther, and this scribe named Gunther confidently states that he has a lifespan of a mean ~quarter-millennium. He can't really explain how he does it in simple terms, but Gunther states that he is an expert at transcribing and storing and retrieving records with great longevity and accuracy.

Gunther suggests that he can keep doing this for as long as he's around, give or take.

Two hundred years ago, I told Gunther to file a transcription of a a 512-byte thing and to store it where he can find it upon my command for as long as he may live. Gunther did so with glee, and stored it however he did.

Today, 200 years hence: I ask Gunther to retrieve that 512-byte thing and email it to me. But Gunther cannot do so; for his perspective, the 512-byte record is gone. Perhaps he can't see it well-enough anymore, or perhaps it has been destroyed, or perhaps it was stolen or maybe even something else. Maybe he can see the record, but the contents just don't make sense to Gunther anymore after all this time.

Gunther is very sorry about this, and I still don't have my 512-byte record retrieved.

Whatever it was: Only Gunther really knew where that 512-byte block of information was stored, and Gunther was the only soul that had access to the area where it was stored, and Gunther is still very much alive to retrieve it. But Gunther still fails to retrieve it. He went to exactly what he knew must be the right spot, and he just couldn't get it done.

Gunther is a failure in this matter. It may have been a stupendous task, but Gunther failed it anyway.

I place a mark on his failure record, and Gunther knows that he earned that mark of failure.

This marked failure will affect Gunther's Mean Time Between Failures, or MTBF.

---

If Gunther is a hard drive, then: The hard drive now has a recorded failure. It may have taken 200+ years, but failure is failure.

(And in a pure and just world, we can chalk that failure up on the MTBF charts for everyone to see exactly what that MTBF has now become for any given Gunther/hard drive.

Except: We also need to know the success rates in order to tabulate that MTBF. Failure alone doesn't get us there.

And in hard drives, nobody publishes good data about success rates. We don't talk about the drives that live forever; they usually get disused and eventually recycled before they properly fail.

IE, we broadly only discuss failure rates (including outliers), and not success rates (we generally send those most-surviving drives off to make new cars out of in some foundry somewhere; we don't wait for them to fail).

Without both measures, a true in-the-field MTBF cannot be established. A percentage-of-death rate without a percentage-of-ultimate-life rate doesn't mean anything in terms of MTBF.

(Which brings us back to those randomly-selected EXOS drives from Seagate, with a rating of 2.5 million hours MTBF.))

You could say so from a consumer perspective, but it is not what the MTBF you quoted measures. Read errors are not considered a hardware failure. Not an opinion.

That's a lot of work to avoid accepting your previous comment was incorrect.

I prepended that chapter with a question: If the data storage hardware cannot read a block that it has stored within itself, then has the hardware not failed?

I then presented practical examples of how I understand failures to behave in data storage devices, such as Gunther.

Can you answer the initial question?

Or are you just here to brow-beat your version of what you believe to be correct -- without the benefit of citation, or logic, or reason?

(Remember, kind HN user: I did not dismiss you at all. I engaged on all cylinders. And I'm here to learn. If you have something to teach, then: Please teach it. If not, then: Please disengage.

I'm not interested in logical fallacies, though, like appeals to authority: Those non-constructs waste our time.)

And MTBF is with active use. Electronics which are not powered on can last almost indefinitely and under a huge range of environmental conditions. From freezing cold to scorching heat.

You just need to make sure you let the electronics acclimatize before powering them on. I always let my external HD acclimatize for 24 hours before powering it on after I grab it from my shed where it's been lying for months in freezing conditions. No problem whatsoever.

Hard drives are way bigger, but M-Disk is more reliable for very long term storage.

The post title is wrong and was poorly transcribed from the article. The product is named M-DISC.

It is a convention that optical discs are discs, and floppy disks or hard disks are disks. It is not a hard-and-fast spelling rule, but since there is indeed a set of official standards for Compact Disc and Digital Versatile* Disc, you can be sure that any conforming optical disc won't be spelled with a "k".

* Were you surprised that the "V" doesn't stand for "Video"?

I use a 2.5" drive. I find them ideal. But then again my backup footprint isn't double-digit terabytes.

Many of the large 2.5" drives are SMR, and I've found sustained write storage go be mediocre at best - 60-80MB/s.

Yes, and it doesn't in the slightest impair their ability to keep my data.

It does impair your ability to transfer it inna reasonable timeframe though! Filling a 4TB drive is a 20 hour endeavour at those rates!

With the decline in quality I am no longer confident my HDDs are actually a save backup strategy.

I have so many photos of vacations and people who I've buried over the years, that I don't want to lose.

So I was thinking about using Blu-ray M-Disks as well, these last couple of months.

Reply on news.ycombinator.com