Music Industry's 1990s Hard Drives Are Dying (arstechnica.com) 259
An anonymous reader quotes a report from Ars Technica: One of the things enterprise storage and destruction company Iron Mountain does is handle the archiving of the media industry's vaults. What it has been seeing lately should be a wake-up call: roughly one-fifth of the hard disk drives dating to the 1990s it was sent are entirely unreadable. Music industry publication Mix spoke with the people in charge of backing up the entertainment industry. The resulting tale is part explainer on how music is so complicated to archive now, part warning about everyone's data stored on spinning disks. "In our line of work, if we discover an inherent problem with a format, it makes sense to let everybody know," Robert Koszela, global director for studio growth and strategic initiatives at Iron Mountain, told Mix. "It may sound like a sales pitch, but it's not; it's a call for action."
Hard drives gained popularity over spooled magnetic tape as digital audio workstations, mixing and editing software, and the perceived downsides of tape, including deterioration from substrate separation and fire. But hard drives present their own archival problems. Standard hard drives were also not designed for long-term archival use. You can almost never decouple the magnetic disks from the reading hardware inside, so that if either fails, the whole drive dies. There are also general computer storage issues, including the separation of samples and finished tracks, or proprietary file formats requiring archival versions of software. Still, Iron Mountain tells Mix that "If the disk platters spin and aren't damaged," it can access the content.
But "if it spins" is becoming a big question mark. Musicians and studios now digging into their archives to remaster tracks often find that drives, even when stored at industry-standard temperature and humidity, have failed in some way, with no partial recovery option available. "It's so sad to see a project come into the studio, a hard drive in a brand-new case with the wrapper and the tags from wherever they bought it still in there," Koszela says. "Next to it is a case with the safety drive in it. Everything's in order. And both of them are bricks." "Optical media rots, magnetic media rots and loses magnetic charge, bearings seize, flash storage loses charge, etc.," writes Hacker News user abracadaniel in a discussion post about the article. "Entropy wins, sometimes much faster than you'd expect."
Hard drives gained popularity over spooled magnetic tape as digital audio workstations, mixing and editing software, and the perceived downsides of tape, including deterioration from substrate separation and fire. But hard drives present their own archival problems. Standard hard drives were also not designed for long-term archival use. You can almost never decouple the magnetic disks from the reading hardware inside, so that if either fails, the whole drive dies. There are also general computer storage issues, including the separation of samples and finished tracks, or proprietary file formats requiring archival versions of software. Still, Iron Mountain tells Mix that "If the disk platters spin and aren't damaged," it can access the content.
But "if it spins" is becoming a big question mark. Musicians and studios now digging into their archives to remaster tracks often find that drives, even when stored at industry-standard temperature and humidity, have failed in some way, with no partial recovery option available. "It's so sad to see a project come into the studio, a hard drive in a brand-new case with the wrapper and the tags from wherever they bought it still in there," Koszela says. "Next to it is a case with the safety drive in it. Everything's in order. And both of them are bricks." "Optical media rots, magnetic media rots and loses magnetic charge, bearings seize, flash storage loses charge, etc.," writes Hacker News user abracadaniel in a discussion post about the article. "Entropy wins, sometimes much faster than you'd expect."
Dual backup (Score:2)
and you don't have to worry as much.
Re:Dual backup (Score:5, Insightful)
You don't gain much if it's all the same age. If it all reaches its end of life, they can easily all be gone.
They only solution is a dynamic backup. One where you recopy it fresh to new media periodically.
Re:Dual backup (Score:5, Insightful)
Re: (Score:2)
Just store the masters in a building that can catch on fire.
Re: (Score:2)
That seemed to work. Untold amounts of content went up in a blaze. Iron Mountain tries to prevent that, but I wonder if they open the drives and attempt to spin the platters in a clean room. This has worked for me, recovering a friend's ancient Miniscribe drive. Saved his bacon, the fool.
Re:Dual backup (Score:5, Insightful)
The 'new' model is planned deletion. Copyright is *supposed* to give the public access to the content at some point. Now with no physical media, there won't be anything to give when it should be public domain.
Which could be described as violating a contract.
TL/DR - pirate everything so it still exists when it should be legally free for everyone to use.
(and also, call your Congress Critters to say to shorten the current period of "75 years after the death of the creator". Like seriously, how is a dead person supposed to be incentivized to create new work?)
Re: (Score:3)
The point is to keep things out of public domain and monetized as long as possible.
Re: Dual backup (Score:2)
Re: (Score:2)
The article literally says frequently the backup (safety drive) is also unreadable.
Re:Dual backup (Score:4, Informative)
That is why you have to test your backups on a regular basis.
The problem here isn't the medium, it's the people handling the archives (or the people above them refusing to give them the resources to do it right.)
Re: (Score:3)
Testing of tens of thousands of drives regularly is a massive undertaking. At some point the expense outweighs the benefits. That's the entire point of the article. If they didn't have to worry about that bit, they could simply back it all up to multiple redundancy settings with numerous means of archiving in different locations and not have this issue. But again, managing and cost is a real consideration at this type of scale.
Re:Dual backup (Score:5, Informative)
Testing of tens of thousands of drives regularly is a massive undertaking.
You don't test tens of thousands. You test ONE drive.
In the 1990s, a one-gigabyte HDD was state-of-the-art.
Today, I can buy a 10-terabyte HDD on Amazon for $100.
So copy the 10,000 1-GB drives onto a single 10-TB HDD, and your job becomes much easier.
That's the entire point of the article.
Yes, but it's a stupid point. The very first lesson of archiving is to periodically upgrade to new media.
They failed to do that and are now blaming their incompetence on "entropy".
Re:Dual backup (Score:5, Insightful)
They can't simply take what's on the old hard drive and copy it to a new one.
1. TFA doesn't say that. If you think it does, please cite the relevant text.
2. There is no technical reason you can't copy from one HDD to another. It is all digital.
They can't take all the drives from the recording of Artist A album 1 and throw it on a drive with Artist K album 26.
Why not? TFA doesn't mention any legal barriers, and there are no technical barriers.
they can't simply take and convert everything to a modern format
Copying a file doesn't require "converting" the format. You just copy it.
You don't just take a simple photo of the Mona Lisa with your smartphone, upload it to Google Drive and claim the backup problem is solved.
That's a silly analogy.
Taking a photo of a painting is a lossy analog to digital process.
Copying a file from one HDD to another is a digital-to-digital process. The new file is an exact copy of the old.
Re:Dual backup (Score:5, Insightful)
This is the universe telling them their copyright is expired. Couldn't happen to be a better lot.
Re: (Score:2)
You didn't read the article and don't seem to understand why they can't just take the original hard drives and copy everything elsewhere.
Re:Dual backup (Score:4, Insightful)
Not any more, because they failed to understand that they needed to copy them every few years. Rather than let them rot for decades.
If your ones and zeroes are gone because of your stupidity and ignorance, they are indeed gone. But while they were still on the drives and readable, they should have been copied. Because new ones and zeroes would have been perfect copies. There would be no difference between them. They would be just as much "original masters" as the original masters.
Because this isn't analogue. Idiotic babble about photograph of Mona Lisa above demonstrates lack of even most basic understanding of this.
Re: (Score:3)
Sounds like they're applying principles developed for analog magnetic tape masters to digital masters.
Which comes back to the same thing: The problem here isn't the media, it's the people maintaining the archives. Either they're incompetent, or their MBA masters are.
Re: (Score:2)
Doesn't anyone else remember those gold Kodak CDR disks [cdrlabs.com] that were supposed to store your data safely for 100 years?
I think I still have some. I wonder if I can find a CD drive to read them on :-O
Re: (Score:2)
Yeah I remember those claims. Wonder what kinda conditions they used to simulate that and make such claims. Clearly it wasn't true.
Re:Dual backup (Score:4, Informative)
Having stuff on another drive is not backup. Remember the 3-2-1 rule:
3 copies of the data,
2 different media,
1 copy offsite.
Didn't do this, they lose.
Re: (Score:3)
If the backup or the backup is an identical drive, then you're not gaining much.
A hard drive + a tape would be better.
Rule of Thumb (Score:3)
I've always heard, rule of thumb, is that you have three copies of your data. An active, copy, a backup, and another backup.
How much are all those thousands of hours of recordings worth? For around a quarter million dollars you can buy a Dell StorageVault with 2PB of space, enough for millions of hours of high res audio, or billions of hours of CD quality audio. Run ZFS on it, then you set up two more in different locations as backups. You leave everything turned on and scrub everything once a month.
It will
Re:Dual backup (Score:5, Insightful)
Way to miss the point.
The drives with the MEDIA, and THE SOFTWARE (which could include applications, licenses, operating system, firmware, etc) become useless.
This has pretty much been the problem since computers moved from tapes and hard drives the size of warehouses. It may be cheap to keep making backups of backups, but who is verifying that that backup can be read? What if it's destroyed by fire? What if it's destroyed because the license to the hardware or software used to play it can't phone home, because that company is out of business, or can not operate the servers any longer (eg Adobe CS2/CS3/CS4 Update servers only work if the local PC clock is set to 2009.)
In the chase down and shake every penny of profit out of customers 'now', leads to a future where that software doesn't work, and the format the software stored things in is impossible to read in other programs.
I actually see a point in the future where a lot of saved pirated media (for example ROM's for arcade machines, and game consoles) is the only way to play the old media (see "Dr.Who" as an example) when the producer has either lost the original master media, and couldn't be arsed to produce master-like versions of the media for general consumption.
It still makes me weep with despair when I see artists flattening their photoshop files into TIFF files to save space. So you have NO desire to ever go back and fix anything in that image if you later decide to print it huh?
Right, now I save streams that I watch just in case youtube nukes the channel.
Re: (Score:3)
Yeah, I get that, you can't play games made 10+ years ago because the software won't work. But I can still run winamp and bring up my old mp3s :)
Re: (Score:3)
Good news is I don't just run windows.
Re: (Score:3)
Re: Dual backup (Score:2)
Micromachines have very little scrap value (Score:5, Funny)
It may be an unpopular opinion (Score:5, Interesting)
I'm tired of all the endless "remastering" anyway. Seems like every time a song gets remastered it just amounts to boosting the bass and applying even more dynamic range compression than they did the last time around. If we're stuck with what was released on CDs to the general public because the masters are gone, well, no great loss. We're not talking scratchy shellac phonographs, so it's not as if people in the future will have to hear some old-timey sounding version of Taylor Swift a century from now.
Heck, by then AI will probably have progressed to the point where you could just ask it to sing whatever song you want as Taylor Swift and the rendition would be absolutely perfect.
Re:It may be an unpopular opinion (Score:5, Insightful)
I'm tired of all the endless "remastering" anyway. Seems like every time a song gets remastered it just amounts to boosting the bass and applying even more dynamic range compression than they did the last time around. If we're stuck with what was released on CDs to the general public because the masters are gone, well, no great loss. We're not talking scratchy shellac phonographs, so it's not as if people in the future will have to hear some old-timey sounding version of Taylor Swift a century from now.
Heck, by then AI will probably have progressed to the point where you could just ask it to sing whatever song you want as Taylor Swift and the rendition would be absolutely perfect.
Most remasters are pretty meh. Some are atrocious. The metal bands are notorious for this. Megadeth's remasters are clearly influenced by 40+ years of loud stages. All the frequencies on the top and bottom that you start losing from all those years at stage volume are boosted while the mids where most of us would prefer the focus to be tend to be dampened. Not to mention the modern loudness nonsense. Leave a little air in a mix and it's beautiful. Smash it into a brick wall and it sounds pretty much like it's been smashed into a brick wall.
Re: (Score:3)
All the frequencies on the top and bottom that you start losing from all those years at stage volume are boosted while the mids where most of us would prefer the focus to be tend to be dampened.
People on average perceive a V shaped sound profile as sounding better so I could see a rando doing this, but you'd expect the pros to have a bit more restraint.
They don't remaster to make 'em sound better (Score:3)
Re: (Score:3)
Who'd they remove from old Ozzy records? like Bark At The Moon etc? FWIW, Black Sabbath songs sound great until Ozzy starts singing. He is no musician.
Re: (Score:3)
Dis
Re: (Score:2)
In some cases they may actually have the original separate tapes or tracks used to record each piece.
Steven Wilson prog remasters (Score:5, Interesting)
In most cases I agree with you, but I have to call out the incredible job Steven Wilson [wikipedia.org] (Porcupine Tree, No-Man, Blackfield) has done remastering prog favorites.
The man takes the original stems and does a complete remix. Philosophically, he's not trying to make an album something different from what it was, but instead to open it up, make it clearer.
I have a couple of his remasters: King Crimson "Red", and Yes "Close to the Edge". In the case of the former, he felt the main mix couldn't be bettered and left it alone, remixed archival "extra" material. It sounds good. In the case of "Close to the Edge" it's an absolute revelation. Sounds completely different, but is still the original performance. It's wonderful.
I chalk up the quality of his remasters to two things:
1. He only works on albums he likes. This means he has a fan's love for the source material.
2. A lot of the masters from the early CD era were just bundled up LP masters sent off to the CD plant. (My original Close to the Edge CD was one of my earliest CD purchases). A bespoke master for digital audio, as well as the latest audio technology and software can make these old tracks shine in a brand new way.
As for a lot of the other stuff - I'm really tired of being encouraged to re-buy, for example, all the old Beatles albums again and again. I don't, unless there's something really interesting there (thinking of "Let It Be, Naked").
Re: (Score:2)
I actually find when I listen to older music it doesn't seem as loud, and that might just be because it's been REMASTERED to be played at a louder volume.
At any rate No, AI will never get to that point. The best you can do right now is RVC and what that effectively lets you do is "remaster" the vocal track using someone else's singing and someone elses voice.
Eg "Sing Umbrella by Rihanna, with the voice of Kyary Pamyu." The result would sound like Kyary Pamyu, but still have the singing style of Rihanna. A l
re: remastering (not wrong, usually sucks) (Score:2)
I have quite a few remastered tracks from various rock songs from the 60's through the 80's and I'd say at best? People would be hard-pressed to know they weren't just listening to the original version. I find a lot of the remastering tends to be pretty subtle and irrelevant to the average listener. Maybe someone listening intently with studio monitors can tell the difference when doing an A/B listening test. But my point is - most of it seems like marketing; an excuse to sell yet another copy of the same m
Wide-Area Distributed backups (Score:2)
Re:Wide-Area Distributed backups (Score:4)
My thought exactly, just give away free copies and you get infinite backups for free.
Re: Wide-Area Distributed backups (Score:5, Funny)
This already exists; it's called BitTorrent, and it's your Friend.
Re: Wide-Area Distributed backups (Score:5, Insightful)
We should actually go back to something more like Napster/Kazaa/Soulseek where it's normal to just share everything you have, rather than just the few torrents you recently downloaded.
Re: (Score:2)
Re: (Score:2)
Not so much now.
Stored Winchester drives as backup? (Score:4, Informative)
That's just silly. These things are mechanical and surprisingly frail. They depend on lubrication and stable mounting. Rarely do the MFM drives of the 1980s work for me when I try to resurrect one. Sometimes. But why should you expect the ATA drives of the 1990s and beyond to be any better?
Old tape cartridges are much better. I have perfectly readable DC2000 tapes from the late 80s/early 90s.
Re: (Score:3)
They should change how the hard-drives are read. Remove the cover and extract the disk. Then scan it on a slow-read turn-table, kind of like a vinyl record. Then you don't get violent head crashes and the like. It takes longer, but gives more options.
Re: (Score:2)
Any competent data recovery service can recover drives in the state discussed, probably at 100%.
But bring your check book.
Re: (Score:3)
It sounds like utter incompetence, but maybe they did not actually want to archive anything and were happy just taking the money for it.
Re: (Score:2)
This is why DLTs exist. We have long known that data stored to tape is less susceptible to failure than on other mediums available.
If the data is important you write it to tape, and verify that you can read back what was written. You store more than one copy on tapes in more than one physical location. Every decade or so, you refresh your backup by reloading it all to a live medium and write it out to a modern tape format again.
This is the way.
Re: (Score:2)
Tape may be better, but it has problems also. Tape can stretch, making it harder to read. There can be print through. (Have you ever listened to a cassette tape and heard a faint version of the first couple of seconds of the recording before the louder regular version kicks in? That's because the magnetized recording affected the blank leader part of the tape which was spooled next to it for a long time when the cassette wasn't in use.)
Also, I don't know if they are still developing new digital tape dri
Re: (Score:2)
Re: They killed DAT before it ever had a chance. (Score:2)
All I can tell is that my DDS1/2/4 tapes from the 1990s are still perfectly readable. The drive from that generation still works. There is still a PCI slot for the SCSI card in my Z170 machine from 2016. OS/2 even boots on it.
Say what you will about DAT, it's far better than hard drives from the 1990s. And CD-R are not archive quality. Many of mine started getting errors well before one decade was up, let alone 3.
Re: (Score:2)
Tape cartridges were engineered to be archival storage.
No, they don't last forever. But they have a much better chance than a cheap AF hard disk. And software makes it very easy to make multiple copies and roll tapes in / out of the library to make sure you don't suffer bit rot.
There's a reason why LTO still exists.
Re: (Score:2)
Honestly, I had kind of poor results with old backup tapes. I think they got established as a solid long-term backup choice in the days of reel to reel tape. But by the time we went to tech like DLT and LTO, or the smaller cartridges like the old Colorado consumer tape drives used? You could easily have a tape trash itself just due to a hardware malfunction in the tape drive (just like your cassettes unwinding in a player, all over the place). Assuming a good working tape drive? They might do better... but
More backups? (Score:3)
I know the article says there is the master then the backup, and both can be bad, but shouldn't there be at least one other copy somewhere else offsite in case something happens to these two? I know it sounds like turtles all the way down, but multiple copies in disparate locations would give a better chance of having a copy in case something happens to the others.
This does bring up the issue of security, but at the very least you've enhanced the survivability of getting data.
Also, even if the drive doesn't spin, aren't there recovery services which disassemble the drive and read from the platters?
Re: More backups? (Score:3)
It *IS* off site.
Iron Mountain is a company that does off-site storage. They used to store paper records in old salt mines, but then moved to electronic records.
The scientific data community usually does a 'media refresh' every 5-10 years. (make a bit-wise copy to new media). It doesn't help with software no longer being usable, but it at least reduces the 'bit rot'. If they get sufficient money, they might also convert files to file formats that are still actively being used (sometimes called 'data res
So, what is the state of the art? (Score:2)
...for archival storage?
Carving it in Stone (Score:3)
Re: (Score:2)
Magnetic tape lasts decades but there are specific formulas of tape that are now failing. If stored properly tapes from the 1960s are still readable.
Re: (Score:2)
Re: (Score:2)
Simple
1.) Archival grade tape that comes with assurances
2.) Spinning disks with regular checks, redundancy and automated re-duplication
3.) Paper (I wish I was kidding).
Option 3. is suitable for example for CA master certificates and other critical low-volume data. Of course you need high-quality laser print or pigmented (non-fading) ink.
There used to be
4.) MOD (Magneto Optical Disks), but the market died because people are too cheap.
Re: (Score:2)
...for archival storage?
Tape. I think we are on LTO-9 now? It is an evolving standard, but still tape.
Re: (Score:2)
The real state of the art is "Make a new copy every year or two. Date each one. Store in a properly controlled environment.". I'm not sure NAND memory will last that long, but disks have no problem...you just have a constant flow of new disks going in.
O, I left out one step. You need to verify the stuff every year or two so that if one drive fails, it doesn't leave you without a good backup.
What media you use depends on what decade you're in, how much data you're backing up, and a few other things. But
Idiotic (Score:2)
A company like Iron Mountain should know hard drives, even in optimal conditions, don't last for more than a few years if not in use. Some of the drives will be recoverable, but only at great expense.
The current best practice for long-term data storage is to transfer to actual archival-quality formats, then re-transfer every decade or so. This avoids rot. (If it were actually archival quality, maybe even longer gaps.)
Once the (practically mythological) storage on crystal is mastered, then archive to that. B
Re: (Score:2)
There is no market for any archival-grade media except archival tape. MODs used to have 50 years assured data lifetime and one Phillips development engineer I talked to said that they were pretty confident for 80 years, and after that the accelerated ageing models broke down, so they did not actually know. I had several MOD drives, very convenient, and the one read failure I had I fixed with blowing dust away with compressed air. But the market was not there, so development stalled and eventually production
Re: (Score:2)
Iron Mountain is just the archive service. The studios own the drives.
Live backups (Score:2)
Offline archives have their place, but to really *know* your data is secured, you need active storage methods. This obviously introduces significant overhead for the long term storage and retention of data, but in cases like this...isn't it kind of worth it?
It's not like we don't know how to do this; backblaze and other online backup companies solved this problem a long time ago. Perhaps it just needs a little tweaking to meet industry specific requirements ( I could see the need for airgapping the backups
Re: (Score:3)
Not really. Archival-grade tape, can reliably be read for its advertised lifetime if stored properly and you have long-term availability of drives. You are correct that the only other reasonable option at this time is permanently checked online storage. It is not more reliable that archival media though, they are roughly on equal footing. What counts in the end is your data retrieval profile and other factors besides reliability.
Re: (Score:2)
They're making this a lot harder than it needs to be.
What they need is an LTO autoloader and a shitload of tapes. I know everyone loves to hate tape, but there's a reason LTO still exists, and it's exactly this.
Re: (Score:2)
Presumably that's why they've hired Iron Mountain to host their data?
Re: (Score:2)
Looks like they're just storing the physical drive, instead of offloading the data to a live backup solution.
morons (Score:2)
Absolutely nothing here is unique to the recording industry, and much of it isn't unique to the particular media either. It's not clear who they intend to deceive with this bullshit, but it's clearly just a grift.
Idiot-Led Effort (Score:5, Insightful)
Tape was designed for archival use, and there is even a special long-term archival variant. The LTO format is backward-compatible at least two generations according to the spec, and often more than that in practice.
Instead of dealing with a slightly slow yet highly reliable technology, they chose something quick and cheap. Anyone with archival experience could have told them it was a bad idea, and I strongly suspect they were told repeatedly.
Corporate execs aren't that brilliant. They deserve neither unchecked authority nor massive paychecks.
Re: (Score:2)
Indeed. This is a massive screw-up and the people that chose this solution are 100% at fault. Any actual data archival expert would immediately have told them that this approach does not work.
Re: (Score:3)
Anyone with archival experience could have told them it was a bad idea, and I strongly suspect they were told repeatedly.
Pretending that actual archivists don't have this issue as well? Congrats you made a tape that is backwards compatible by 2 generations. You just bought yourself *looks up LTO releases* 7 years? Shit man my HDDs last longer than that.
Even archival tape has issues long term. Binders releasing, tapes getting demagnetised, and that's before you consider the technology changes. You say LTO tape as if that isn't something that is over a decade *newer* than archival HDD storage from the early digital age.
But plea
U2 was way ahead on this (Score:3)
Funny enough, it was almost exactly 10 years when U2 decided to backup one of their albums onto everyone's iPhone. Making lots of copies everywhere seems like a great solution.
https://www.stereogum.com/2279... [stereogum.com]
Re: (Score:2)
U2 on iPhones was really 10 years ago? Flappy bird [slashdot.org] 10 years ago? It seems like Slashdot is dedicated to making me feel old today!
Re: (Score:2)
I've been lurking on slashdot since before the turn of the century, which makes me feel old.
Effectively permanent solution (Score:2)
Etch your audio into a quartzite disk as if it were a vinyl record. Excluding gross mechanical damage, it will remain playable until the planet is melted by the expanding Sun. Longer if you ship it beyond Mars' orbit.
Not just the music industry (Score:4, Interesting)
My father worked as a "sensitivity reviewer" for a UK government records office in his retirement after he left the diplomatic service. The stuff he was releasing was from mostly 50 years years ago when it was statutorily made public as long as it didn't deal with anyone alive (or members of the royal family - thanks Tony Blair!). He said he'd heard that there was a "black hole" coming down the line from the period in 90's when departments moved to email and digital storage. People in about 2010 trying to get stuff off early archives of those for internal purposes and FOI requests were reporting they were often unable to do so.
Maybe that's a good thing, I dunno. But for anyone wanting to know what was being said about, for example, Pergau Dam in internal communications between Whitehall and the Foreign Office - they might have a hard time.
just revive piratebay (Score:2)
If it spins ... (Score:2)
But "if it spins" is becoming a big question mark.
One of my PCs in the late 90's ran Windows NT 4.0 and had a 5GB F/W SCSI hard drive (and SCSI CD-ROM), it ran 24/7 for over 5 years, then sat around in a box for 5 years. Before I finally got rid of it I wanted to double check if there was anything to copy off. The drive wouldn't spin up, and just made regular soft clicking sounds, so I rapped the side of it with a screwdriver handle and it then slowly came up to speed. I let it idle for a while then browsed the drive, all the data was fine. Luck? Pro
Re: (Score:3)
Re: (Score:3)
Not luck, it is well known that the lubricant in the bearings thicken up to the point the motor is unable to get it spinning without help. Usually it involves some light vibration as you did or sticking it in a freezer to make the lubricant shrink and fragment.
Interesting I had success with the opposite approach by putting heat energy into the drive. Instead of vibrating the drives I'd use an oven to warm up an old drive and that would often overcome the Stiction [wikipedia.org] that stops the disk from spinning.
What I do for stuff I've recorded and mastered is regularly transfer stuff off older drives which is better than putting drives in ovens to recover data whilst reducing the physical storage space required. I'm a little surprised that Iron Mountain don't already do t
Re: (Score:2)
This was a startup procedure for old OS/2 shit at a place I worked at almost 20 years ago. Open up the chassis and tap on the 9.1GB SCSI drive with a tool while applying power. Then start up a process to create a disk image of that piece of shit because it's done, and while it's copying everything off, go find a new drive to restore it to.
Re: (Score:2)
Sounds like a Quantum.
They archive things on non-spinning HDDs? (Score:2)
How incompetent can you be? That this does not work has been clear, basically forever. Who are these jokers?
Your choices essentially are 1) archive-grade tape with actual verified assurances on how long they keep data and 2) archive on spinning disks and do regular checks and have enough redundancy and fault-tolerance. There used to be 3) use MODs, but nobody cared enough for an affordable archival-grade medium with >50 years data life expectancy.
Re: (Score:3)
It amazes me that people just completely forgot that LTO is still a thing, and the #1 reason for it still being a thing isn't because of old neckbeards that refuse to let dying crap die, it's because those neckbeards actually know what the hell it's good for, and what it was designed to do. And autoloaders scale quite well, especially if there's more than one drive in them.
And those neckbeards really hate being called at 2am because someone fucked something up and the backups are no good.
Re: (Score:2)
Exactly. And it is not like this is a legacy technology that does not get improved anymore. This is current, cutting-edge tech, for when you know your stuff and need it. Obviously, it is not cheap. But it scales up to whatever size you need.
I guess this is one of the things where actual IT experts and all the wannabees separate themselves into two groups that are really very different.
When have hard disks EVER been archival grade? (Score:4, Insightful)
Pardon the word capitalization, but when have hard disks been ever considered something of an archival grade medium. At best, it is something to take offsite for 3-2-1-1-0 (three copies, two on different media, one offsite, one offline, zero errors) backups, in a grandfather-father-son rotation.
The problem we have had in the past decade is the fact that tape has become so much more expensive than hard disks. It used to be that one could buy a tape drive for a few hundred dollars, or a really good SCSI one for a thousand or so. Now, if you want something with any relevant capacity, you are looking at $6000 just for the drive. Similar with optical, where CD-R disks and DVD+/-R disks could be relevant to system backups, and even Blu-Ray media. Now, BDXL only goes up to 100 gigs, which is relatively tiny, because $100 can get one a 5 TB removable USB hard disk... yes, it will be a 2.5" SMR drive, but that isn't bad for throwing data on to take offsite. Of course, one can spend around $70 for a much larger, refurbished drive and stick that in a case, and with a USB adapter, it will be faster and CMR.
Problem is that drives die. You keep them in ideal conditions, they will have parts break, or because of the push for ever higher aerial densities, this sacrifices lifespan. The little bearings wear out, causing things to crash, and once the servo tracks are not readable, the drive will just tick forever.
What we need is to get some companies to start working on newer storage formats. Of course, holographic storage has been in the wings for decades, since Tamarak in 1991, Inphase in ~2008 (who eventually had all their IP bought by Apple). So, we have tape, which can see a good amount of aerial density improvements, just because it had a lot more space to put data on than a HDD platter. 3148 feet by ~1/4 to 1/2 inch (doing rough calculations in US units gives ~10,000 square inches of area, compared to ~75 square inches of a HDD with multiple platters and both sides.) I'm sure there is a lot of improvement that tape can get, even factoring in that tape has to physically move and get wrapped and unwrapped constantly.
Optical is another. There are a ton of technologies we can use to throw more types of bits onto burnable media, not just layers, where you can increase the layers (China has a 100 layer optical disk), use higher frequency UV, use different shapes of pits to allow more than zero and one states (similar to TLC and MLC cells on SSDs), narrower spacing between tracks. Once done, optical is relatively cheap, and the technology is mature. It is just surprising we have not seen any optical upgrades since BDXL, because this is the ideal media for long term storage [1]. It is WORM based (well unless one does go for RW media), relatively cheap, easily stored, and because the data is stored mechanically, magnets won't affect it.
On a personal level, I've pulled out CD-Rs I've made in the 1990s, and because I used WinRAR, was able to verify that even after these many years, all data stored is intact.
[1]: No media is 100%. Ideally, every few times a year, media should be checked for errors, and if there are some, another disk generated to replace the missing one. Goes without saying to have multiple copies, perhaps in different areas.
for storage reasons alone. (Score:2)
This is negligence (Score:3)
It's unsurprising that 25 year old hard drives are failing. It's surprising, though, that no one's done anything about it before now. Audio files are not large. Hard drives in 1999 were, what, 30-40GB at most?
Current state of the art for backup is LTO-9 tape, which is 18TB to a tape, and the tape costs around a hundred bucks (less in bulk). That means 600 90's hard drives could be backed up to a single LTO-9 tape. The most expensive part of the process would probably be connecting the hard drives and copying data off of them, but once you've done that, it's easy to make multiple LTO copies for essentially no cost, which lets you store them in different locations, etc., and the fact that they'll take up maybe one thousandth of the physical space/weight has to lead to pretty good savings over time.
LTO tapes have at least a 30 year lifespan, and they don't fail all at once like disk drives do, so just plan to copy the data onto new (larger capacity) LTO tapes in 15-20 years, and you're golden. That no one has thought of this for 25 years, and just allowed these hard drives to sit around and rot, is pure negligence.
Spinning disks (Score:2)
The resulting tale is part explainer on how music is so complicated to archive now, part warning about everyone's data stored on spinning disks
No, this has nothing to do with spinning disks. It has to do with trusting any single medium. Magnetic tape, SSD, CD/DVD...they *all* have a shelf life, and can be damaged in various ways. If something *must* be archived long-term, you have to keep it "alive"...transfer it from one medium to another over time. There is no storage medium that you can just leave alone and expect it to still be readable a long time in the future. For digital formats, it's even worse, because even if the physical media are stil
Re: (Score:2)
Another thing. It's possible to put stuff you know in the future can be used against someone because it'll be in the data .. sort of like blackmail.
Re: (Score:2)
Sometimes wonder if storing all this data about our past is a good thing? One example, in the future if people become vegan .. they may look at us who eat meat the same way people think of slave owners.
Are you suggesting that future generations are going to be judging our society based on disco or something? "Your ancestors listened to the Bee Gees, they were monsters!"
Re: (Score:2)
in the future if people become vegan .. they may look at us who eat meat the same way people think of slave owners. Proven descendants of people who used to eat meat would get discriminated against, meanwhile the people who have no records of what their ancestors were like will proclaim all their ancestors were vegan.
You win for most random comment lmao.
Re: (Score:2)
So, I've got a bit of practical experience with this, and uh... it would sound crazy if I said why or how, so we'll just leave that part out. What I can tell you for sure though is that in the future they'll look least kindly on the periods who either purposefully destroyed or failed to preserve their own history for posterity, and they'll look most kindly on the periods where the most data was preserved, even if lots of it would have seemed unnecessary or frivolous at the time. By and large, ethical differ
Re: (Score:2)
“By and large, ethical differences will be taken in context and not really judged critically by current standards as you expect.”
Tell that to the people trying to use the fact that George Washington held slaves for various agendas. I mean, I get what you’re saying though. I would love to know what daily life in Egypt was really like. What did people talk about, joke about etc? Someday we’ll be the ancient Egypt. Who knows, maybe some super intelligent AI augmented descendent will rea
Re: (Score:2)
Welll, there are two ways that are durable. Both are quite inconvenient.
1) Paper tape. (But you've got to be careful not to depend on a plastic layer in between the layers of paper.
2) Glass disks with a metal layer, into which pits are burned with a laser.
2 it the thing that developed into the CDROM, but the original form was a lot more stable. It was also a real pain to write. (I don't think it was ever used by even a small company, only by rather large ones that had a special need.) Note that this pr
Re: (Score:3)