Really the one thing that's guaranteed to be manufactured in N years is _some kind of storage_. I wouldn't buy LTO-1 now, but later gens are going to be around for some time, just like CD-ROMs and other optical media, in some form. And readers will always exist barring an asteroid impact or whatever.
>Therefore I already know the answer on whether I could possibly buy newly manufactured LTO drives and media in 10,20 years, and it starts with a NO. Even today I would be forced to buy second-hand drives. Why would I even entertain LTO as an option, then?
This is just yet another example of a pointless argument, exactly as I mentioned. Why should I even care that you can scavenge for ancient stock (even if NOS), when my current concern is about something that is still manufactured _today_ ?
I know for sure optical media & DRIVES will still be available to purchase _brand new_ during the N years they're still manufactured, but also the M years that will follow where I will be able to find new/old stock after they stop manufacturing.
Period N by itself I expect is going to be somewhat long (see 3.5inch floppies), during which one can even expect to see drives with never interfaces (e.g. USB-C). Yes, I have no clue how long it is really going to be, and my concern is whether it will even last this decade.
OTOH I know 100% for sure period N is going to be effectively 0 for any LTO generation I could possibly buy. By the time LTO prices drop for some generation, it is because that generation is dead in the water.
And period M? It is going to incredibly long for optical due to popularity alone, much longer than _any generation_ of LTO could ever hope to be.
And if you say "well, certainly some form of LTO is going to be manufactured in 20 years from now": it should be obvious that I couldn't care less, unless that form of LTO would be able to read the tapes from any generation I can possibly buy now.
The fact that LTO-21 will still be manufactured is of absolutely no relief to someone with LTO-4 tapes. In fact, for all I'm concerned, it could very well be an entirely different media type only sharing the first three letters of the name.
These are not arguments in favor of LTO. If you're already assuming that if your LTO drive breaks you either scavenge for another or basically assume the loss and buy all new media from newer generation and repeat... what's the point of LTO then? Why not buy SmartMedia cards (to say the worst thing that comes to mind)? I'm sure you can scavenge readers and media, and probably will have an easier time finding and using them than with any specific LTO generation.
In the meanwhile, let me keep burning toasters; at least there is a small chance I may be able to buy new drives 20 years from now, using whatever interface replaces USB-C, and they will still be able to read my current discs.
I mean you said it yourself: if you have terabytes of data, BD isn't practical.
I feel like this is all just two totally separate use cases. Nobody wants to burn 20-40 BDs per TB, just like nobody wants to use a tape drive (or maintain a RAID array, or whatever else) to back up 500GBs of family photos and tax documents or whatever.
At some point the volume of data dictates what solutions are practical.
This is true - I got a fiber channel LTO-8 FH drive off ebay brand new in the IBM packaging for less than 750$ Tapes are 60; so breaking even against 15$ per TB HDDs is pretty fast.
I believe they are all 5.25in, some are just in a case. Even the library drives are just two 5.25 bays put together, a full height drive; vs. the much more common half height.
You can still buy brand new LTO-4 and up from a brief search - I think due to the enterprise use cases it’ll hang around longer than any other format. Tape existed before the HDD; it’ll be there watching HDDS pass away into the ether too. Probably a few tape drives on the Starship Enterprise somewhere.
More seriously; you can buy used lto-7/8 for very little these days, and the tapes are extremely cheap per gb. The drives are somewhat loud; it’s not a beside device for sure. I’m finding it a bit of a pain to manage a good backup strategy with them.
- You suggest buying multidecade old drives that are no longer manufactured, have weird interfaces that your 2026 PC no longer has, are expensive, large, noisy
- You then mention LTO7 which will not read your LTO4 tapes and is not just expensive but literally out of reach economically for single home
Basically LTO is a terrible backup strategy unless you have a lot of money regularly that you will spend in order to upgrade your entire equipment every two/three generations (otherwise your newer equipment wont read your old tapes). Or you have so much data to backup that cost of drives is not really an issue.
Using HDDs for backup is also a terrible long term strategy, because you must have a lot of money regularly, to buy new HDDs to replace your old HDDs and this much more often than you need to buy a tape drive to migrate your tapes.
I have stored a lot of data on HDDs, and the only reason why I have not lost any of it yet is because I have always used duplicate HDDs. After 5 years or more, most HDDs had some corrupted sectors, but they were not in the same positions in the duplicate HDDs, allowing complete recovery of the data.
The reality is that both tapes and HDDs suck. What is really needed for long-term storage is a write-once memory with a lifetime of 100 years or more, based on an open standard that would ensure the availability of readers in the future.
If such a memory would use optical reading, it would have to use a great number of layers, filling a 3D volume, in order to achieve densities comparable with the magnetic media. While several research projects in this direction have been announced from time to time, until now none of them has resulted in a commercial product.
HDD cost small dollars only for small amounts of archived data, i.e. up to 100 TB or 200 TB at most.
For greater amounts of data, HDDs become too expensive and this is the main reason to switch to tapes.
Obviously, for someone who is certain of never needing more than a few tens of TB of storage space it would be foolish to use LTO.
On the other hand, for someone storing 500 TB, it is foolish to use HDDs, because tapes are more reliable, more compact, faster for sequential transfers, i.e. the actual backup and restoring, and cheaper.
It is as simple as that. The decision of using HDDs or LTO is strictly determined by the amount of data that must be stored.
The argument that HDDs should be fine for most non-technical people is correct only because those people do not store much data.
I'm confused - I and many others basically have these cobbled together LTO setups. I'm only "Prepping" by finally moving some of my backups away from home, so in case of a fire or whatever I'm not out of luck. You could cobble one together now for anything from OG DAT tape to LTO-10 for ~10K, if you need. So big fire happens, you file an insurance claim, and as part of the payout buy whatever setup you need, or hire some specialists. Once we are at LTO-20, there's no reason to assume LTO-10 and older drives are totally gone from the used market?
I'm not preparing for some asteroid impact level event, in that case the loss of my backups will presumably not really matter all that much.
That’s babble to 80% of the nerdy HN audience. “Copy your stuff to this usb drive and keep it somewhere you aren’t” is easy for almost anyone to comprehend, accessible but operationally difficult.
Maybe you're having issues with their writing style or something but the tech is simple. They copy their stuff to a tape and keep it somewhere they aren't. If a disaster happens they'll buy a new tape drive.
I bought a thunderbolt to FC adapter; works perfectly on Mac and Linux.
I mention LTO 4 because you can today, buy multi decades old LTO-4. Brand new. So in multiple decades from now, I assume you’ll be able to find LTO-7 or 8; brand new. A drive might cost a little more to obtain, but given the plethora of used multi decades old lto currently out there, it seems reasonable to expect that in a recovery scenario you’ll be able to shell out for the right drive.
But yes for most HDDs or the cloud are better. No need to get spicy about it.
I'm not going to actually suggest LTO-7, but what do you think is a reasonable per-month cost for backing up your important data? If it's in the $5-$10 range then you can afford a $600 drive and some tapes.
> Basically LTO is a terrible backup strategy unless you have a lot of money regularly that you will spend in order to upgrade your entire equipment every two/three generations (otherwise your newer equipment wont read your old tapes).
"regularly" can be 10 years. Your new equipment doesn't need to read your old tapes. If you advance by 4 generations, you can buy 1 new tape to replace 10 old tapes. And the newer generations have abandoned that feature anyway.
AFAIK the tapes are cheap, but tape libs aren't. Considering that they also take up a significant amount of space, I personally don't see them as a viable backup medium for most private users.
You don't necessarily need a lib, though. Especially if you're interested in a use case where you can store data in a go bag, safe deposit box, etc., it seems like having individual tapes would be preferable.
Individual used drives aren't too expensive (or at least didn't used to be). Libraries, in contrast, do tend to be more expensive (and also a lot more trouble to ship).
“We believe we have found the cause of the problem and are working on mitigation.” Wow - they believe they’ve found it! Comforting.
Having worked at Microsoft in the recent past I’m not surprised everything is starting to fall over a lot more.
While there are tons of good people at Microsoft, the overall culture and number of people who stand in the way of getting stuff done well is immense. People view you or your team doing work as an existential threat to their team, and react accordingly.
I don’t think the recent uptick in failures is related to vibe coding or AI and has everything to do with the corporate culture. Probably some poor product manager was tasked with closing 10k automated security flags/tickets and in doing so broke some load bearing microservice that depended on a kind of service principal that’s now not allowed.
Wow you know it’s a fun party when the first result needs to specify it’s _not_ radioactive
Oh I see - Uranyl Acetate is radioactive and this replaces it. Fun!
This seems like a friendly chemical too -
“ The chemical properties of Osmium Tetroxide are such that use and handling of the chemical is often considered daunting. Although its volatility and toxicity certainly makes it a dangerous chemical, but when following the proper procedure and taking the necessary precautions, Osmium Tetroxide can be used to its full potential with limited risk to the user.
This is more toxic than glutaraldehyde and has a higher vapour pressure. Particular care must be taken to avoid breathing the vapour or allowing it to affect the eyes. ”
Uranyl acetate for staining is typically depleted and unless you have regulatory issues I don't think the radiation is a big concern, especially when you compare to the very serious toxicity of OsO4 (vapors can react with your eyes and blind you).
Interesting and makes sense! I know nothing but what I read from the stain description haha. OsO4 seems incredibly nasty. So do a few other of the stains!
I believe the author is talking about the OCP (2.0) network card itself, that these adapters internally. The OCP nics are quite cheap compared to pcie - here’s 100GBE for 100!
https://ebay.us/m/HMQAph
This 100GbE card is an OCP 2.0 type 2 adapter, which will _probably_ not work with the PX PCB since that NIC has two of these mezzanine connectors, and PX only one.
What also may not work are Dell rNDC cards. They look like they have OCP 2.0 type 1 connectors, but may not quite fit (please correct me if I'm wrong). They do however have a nice cooling solution, which could be retrofitted to one of the OCP 2.0 cards.
I've also ordered a Chelsio T6225-OCP cards out of curiosity. These should fit in the PX adapter but require a 3rd-party driver on macOS (which then supports jumbo frames, etc.)
What also fits physically is a Broadcom BCM957304M3040C, but there are no drivers on macOS, and I couldn't get the firmware updated on Linux either.
That’s a good point to note! I think the stacking height would matter, but in theory the single connector is still 8x pcie and should link without the upper 8x lanes connected.
>Therefore I already know the answer on whether I could possibly buy newly manufactured LTO drives and media in 10,20 years, and it starts with a NO. Even today I would be forced to buy second-hand drives. Why would I even entertain LTO as an option, then?
You can still buy brand new LTO-1 media from 2000 - 26 years old. You shouldn't, but you can. https://www.malelo.com/Maxell_LTO_1_Ultrium_Tape_100_200GB_1... Then here's a cheap drive https://www.ebay.com/itm/355784908408
So if you needed to restore a backup from 26 years ago, it would not cost you very much.
reply