hey it's that time again.

do you have data? that you like? is it backed up? do you have automatic backups you've checked are actually happening?

in multiple locations?

have you done a test restore to make sure the backups work?

*even an extra copy on a flash drive is better than no backups*

reminder that flash storage is not long term storage
@gloriouscow If you’re planning long-term enough that flash data retention is an issue, then it’s not like spinning drives are long-term storage either. The only substantial step up from flash is a live system which you regularly scrub and service.

@bob_zim there was no intent behind it perhaps but i can't help but think about how much data vital to computing history was recovered just because someone found a shoebox full of hard drives

if someone finds a box of m2 drives in 40 years they will just be unreadable e-waste

like if you chart the amount of "unintentionally preserved history" over time it will just plummet in a corresponding line to the adoption of flash storage.

that's kind of sad :(

@gloriouscow Most current hard drives will also be bricks over that time span. And most tapes. Outliers will exist, but that’s survivorship bias.

I think the prevalence of encryption is also likely to reduce the recoverable data from today.

@bob_zim @gloriouscow Spinning drives do do better in the medium term, especially if they remain powered-on. As always, multiple copies with ECC or hash-check is wise.

@shelldozer @bob_zim

the miniscribe hard drive in this IBM 5170 I bought spun right up and I was able to read this guy's creepy letters about his ex wife.

actually this is not a good argument for data preservation i'm sorry

@dalias @gloriouscow Even tapes don’t survive well unattended. The base gets brittle over time and fails. They last better than flash or spinning drives, but not *that much* better. I’ve been asked to digitize 30-year-old VHS and reel-to-reel tapes which just crumbled when I tried to examine them.

Long-term data preservation takes active work. If you want something to last decades, you need to build it to be fault-tolerant, you need to test it regularly to catch failures before they exceed the fault tolerance, and you need to fix the failures.

@bob_zim @gloriouscow That's really dependent on storage climate. If you don't have high moisture and fluctuating temperatures, they'll last a very long time.

@dalias @gloriouscow SSDs stored in a controlled climate will generally also also retain data for decades. My point is the data loss concern for unattended flash is overblown (much like concern about flash wear), and other media types are also risky to leave unattended.

Basically anything but burnable optical media will last at least two years stored in a climate which doesn’t kill people. Store it with climate control, and it’ll last a decade, maybe two. Tape might last twice as long as an SSD, but it won’t reliably make it to the next bracket up.

@bob_zim @gloriouscow A few weeks ago I read back hundreds of GBs of burnable optical media that hadn't been touched in 20 years, with no errors. Stored in hot attic for half that time. 🤷
@gloriouscow Why? Technology wise, I know you're supposed to power the drive on from time to time to keep the charges from degrading, but every time a storage medium's failed me, it was always either a cheap USB flash drive or HDD. The first SSD I ever bought is still alive.
@NanoRaptor i read "do you have a date" and the more i read on the more confused i got.
@yetzt i have multiple tall muscular firefighter lasses with ukrainian accents stashed around the place. and i test them regularly!
@NanoRaptor counterpoint: I have 50TB of data. I can't afford to back it up
@freya similar tbh... but i do have a core of irreplaceable ones. My NAS is a little one, its backup drives little too, relatively.
@freya @NanoRaptor
Can you afford to lose it?
@kirtai @NanoRaptor I am on SSI. I do not have the money to dedicate to backing it up. it's not a tradeoff, it's a literal state
@NanoRaptor Backups? In this economy??

@NanoRaptor I've got a daily backup to a USB external drive, a weekly backup to a NAS, a monthly-ish backup to a second external drive that's disconnected and stored in a more secure corner of the house when not being written to, and a once or twice a year backup to *another* external drive that I store at the house of a relative who lives in a different town. Plus another two full copies of my system on external drives intended less as a backup and more as a way of having the ability to get at anything I could possibly need while on my laptop, and another 2-3 portable SSDs with select items.

It's actually kind of a problem...

@lusrangifer if i'd hazard a guess i'd say you too have once lost everything due to no backups way back in the past...

@NanoRaptor I actually have not!

...but during my university days I worked part-time in the wing of the IT department that offered tech support to students, and that campus was rough on laptop hard drives.

@lusrangifer Oh that's grand then!

I lost everything in 2000, and only got it (mostly) back more than a decade later.

Still have The Fear!

@NanoRaptor
I have all my data backed up twice! Because it's in RAID, sitting next to my computer. Right?
@NanoRaptor - but remember to plug in and use the flash drives or SSDs every few months - I am seeing data corruption and loss on SSDs that I haven't used for only about a year. Spinning rust hard drives fare better than SSDs!
@NanoRaptor daily backups to a couple of USB hard drives I can grab and run with if the apartment’s on fire. Weekly backups to Backblaze, soon to be a 16tb drive sitting in a DC halfway across the world. I’d like to add an extra offsite copy when disk prices aren’t stupid.
@NanoRaptor I don't like it. I hope the disk fails and we can finally be free.
@ellie there is some aspect of this gods, yes.
@NanoRaptor "But Bob probably backed up his work and Sue backs up the shared folders. And the trainee said he sometimes backs up stuff too. You're just trying to sell us an expensive backup solution that we don't need."
@NanoRaptor My backups are great and tested tyvm (https://www.alkoclick.space/my-backup-strategy), but maybe some of the storage nerds in this thread can help me: How can I estimate bit rot on a modern SSD (I'll take any model) over a twenty year period of roughly ideal circumstances? I've been searching all day and I'm now down to academic papers proposing arcane optimization algorithms and still I fail. I must be missing something
My backup strategy

I am writing this document for myself, as well as anyone else that’s interested in taking a stab at backing up their homelab tools.

🔮 Enchanted Systems

@NanoRaptor Answering my own question now that I understand the topic a bit better: The closest number to what I had in mind is UBER, Uncorrectable Bit Error Rate and for most modern HDDs and SSD is somewhere between 10^-13 and 10^-16. In practice the actual Raw Bit Error Rate on operations can be as low as 1/1000, but SSDs use a bunch of Error Correction Codes to validate and fix data as they read it, so you end up with the UBER number.

As a fun extra. the question as posed is kinda invalid: SSD longevity, even under ideal circumstances is heavily reliant on the load of read/write operations, so basically the amount of ops translates to SSD age, unlike HDDs.

Refs used:
* Data Longevity and Compatibility
* Data Retention in MLC NAND Flash Memory: Characterization, Optimization, and Recovery

https://web.ece.ucsb.edu/~parhami/pubs_folder/parh19f-ebdt-data-longevity-compatib.pdf

https://users.ece.cmu.edu/~omutlu/pub/flash-memory-data-retention_hpca15.pdf

This and other things about SSDs here: https://www.alkoclick.space/things-i-learned-about-ssds

@NanoRaptor also: do you have a way to get to that backup if you don't have access to what that backup holds? Any logins won't be helpful in restoring from backup if your password db is only in the backup. And 2FA can also get in the way if you're not careful.

@NanoRaptor

These are the correct questions in a nutshell.

@NanoRaptor
It's fine Google says they got me covered, even though it's a free account. 🙃

@NanoRaptor I store all my critical data* in the Fediverse!

*cat memes

@NanoRaptor in order:

yes yes yes and yes!

yes, but not enough. fixing that this weekend was already on my docket (for things arent local mirrors of easily available data, anyway)

yes, im actually running my home storage on the live backup as i write this (with "prod" as backup meanwhile) and will be switching back over to "prod" this weekend

@NanoRaptor

Four mirrored RAID arrays, everything copied to an internal disk, an external USB drive, and a NAS, plus cloud backup. Yes, I'm paranoid. 😆

@NanoRaptor
I swallowed my flash drive and now I’m backed up.

@NanoRaptor Glad when I switched the last bastion of Windows, my home desktop, to Fedora recently, I did so while doing a storage upgrade/expansion.

Installing Fedora on #ZFS was a challenge, but thanks to ZFSBootMenu it worked after figuring stuff out, and thanks to syncoid/sanoid I now have incremental and full backups that get synced daily.

Still have to find a solution for offsite part of the 3-2-1 rule that does not break the bank.

@NanoRaptor darkling's first law: the user hates their data.
@NanoRaptor Good reminder, always 👍
@NanoRaptor Yes, it's that time again...
@NanoRaptor I want to try blueray M disks.
@NanoRaptor
Just had a disaster a few days ago and I'm so f glad I stored an extra copy in a different location.
Triple copied across my home server, desktop, laptop, and partially on my smartphone. Had to do backup restores already.
@NanoRaptor 😅 uhh yeah about that...
@NanoRaptor it's not that we don't want to, it's that we're very dysfunctional when it comes to that

@elexia yes! and this is why i'll alway include just dumping stuff to a flash drive counts cos for some of us setting up more than that is functionally REALLY hard!

I lost everything in 2000 and it took until *2016* until I set up truly solid backups.

@NanoRaptor remember back in 2006 or so, our branch failed to pass backup policy checkup, because weekly incremental backups haven’t been stored in bank safe storage in distance of at least 5 kilometres away from our offices. It was only 4.7 km. All the other metrics we passed. We did had every 6 month drill to check our backups. In theory CTO randomly selected time and file/dataset and we needed to make proof we have backup from that point of time. Remember one day he had selected one particular email. We needed to spin up secondary instance of ms exchange server, restore whole mailboxes dataset and copy out that particular email.
@NanoRaptor probably the first time in my life that I can say yes for the last 3 month... sigh.. I am still avoiding to figure out if I may have lost something very precious in the last transition T_T
@NanoRaptor I miss being able to back up your entire computer on maybe 1 or 2 cheep optical discs. That said, i'm still using Bluray for extra backups
@NanoRaptor Offsite encrypted backups in 4 different locations and 4 different media.
I'm beginning to think I'm overdoing it...

@NanoRaptor
Good post.

I have a Diskstation (called Daisy) with twin RAID HDDs.
I have a Diskstation (called Daphne) with a single HDD.
I have a Pi 5 (called Norah) with an NVMe on a hat.

All my vital data heads to all 3 of these via Rsync on a daily basis and I check it from time to time (no failures yet).

And I just got some Proton cloud storage so I'll send some stuff there too.