Hacker News Reader

Comment by Aachen | Hacker News Reader

I keep telling people to make backups of "cloud" services ever since Grooveshark died from one day to the next

Maybe Spotify won't be taken down overnight, but they can lock me out for various reasons such as misdetecting an IP address (Google wouldn't let me log in at all when they thought I was in Russia, when I was at 31C3 in Hamburg, Germany, a conference where they afaik use some temporary IP space — I don't know what consequences that would have nowadays)

Hacker News also: if there are things here you want to keep, store them yourself

Facebook chats, Signal, etc.: download the data in a format you can decode if there's anything you want to keep. Make sure you can decode this Signal backup format if your phone dies

View on HN

> I keep telling people to make backups of "cloud" services

The usual rule when making backups is that you need 3 copies, one of them offsite. A cloud storage service counts as an offsite copy, but you still need two more, typically one on your computer and one in cold storage (ex: external hard drive).

Treat cloud storage like you treat your hard drive, it can fail at any time. The failure modes are different: mechanical failure vs losing access to your account, but the end result is the same: the system is unreliable and you have to engineer around it.

I think applying these traditional schemes to SaaS environments (now called cloud) can be misleading

When people want off-site backups, the server owner can say they've already taken care of that. I'd say that having a copy under your own control is a separate checkbox we should include!

Also backups in two separate legalizations if the hosting company is not in the same country as you or your company.

Good advice. Though I’d say you can count the cloud storage as 2 copies. They typically replicate the data to three drives, that covers all the typical failure modes you’d normally get when running a drive yourself. Then with a simple click you can even get a replica in another region.

The third backup being in a decoupled system on site is what you now count as your ”offsite” backup. Having this last copy is non-negotiable, no matter how many 9s S3 claim to have.

Yes and don't forget signal is rolling out free cloud backup for texts so it should be making the process of backups even easier for non techy users.

How do I get it out of their cloud? It works as a backup to my phone, but I'd still want another copy that I can access regardless of whether I've got access to that phone number right now. Sorting that out with the carrier can take a while

What's (to me) much more exciting to me is the incremental local backups they're working on. I've iirc looked into how to decode these big backup files, and copy them off my phone semi-regularly, but this being rsyncable will make the process much faster

I assume the process is basically to restore from cloud on device and then export a local backup. But I don't recall them saying anything specific about that. The incremental backups do sound really nice.

That sounds like it'll require access to the phone number to set it up on a new device, and that's assuming the user promptly has the funds for a new device

Even if it works for desktop clients also, that still needs to be set up using a data matrix scanned by the mobile device (or if that's already done before you lost the origin device, you can "simply" extract the faux-encrypted sqlite format that the desktop client uses, albeit with having to recompile this encryption extension for Sqlite manually — ask me how I know..)

I'm still heartbroken about Grooveshark all these years later.

I still have my copy of [SciLor's Grooveshark.com Downloader][1] for posterity's sake :)

[1]: http://www.scilor.com/grooveshark-downloader.html

> even with Grooveshark being down it still works!

Wait what, how does (did?) that work?

And more importantly (I can find the opus files myself): can/could it retrieve my account data to know which songs to download?

I've been able to partially reconstruct my music from localStorage, downloaded songs on my phone (yay for having root and access to that data folder), and memory, but it's still incomplete

Fwiw, here's my own little contribution to getting song data from Grooveshark, in my first blog post: https://lgms.nl/blog-1 (not my first blog site though). Happy to meet other fans :D

Any good tools for automatic backup of these things? How would a non-tech-type do it?

This probably doesn't qualify as a non-tech type answer, but I have a Synology NAS which includes a feature (CloudSync?) to automatically download Google Drive documents in docx, xlsx, etc. format.

Ooh - I have a Synology NAS and didn't realize it can convert files on sync.

It's also ironic that we're considering docx etc as open formats these days.

There's sadly no silver bullet, especially for non-techies. To me it's like buying vegan products where available: I can't trivially become vegan with no effort, but I can affect market forces by preferentially buying things that align with that ideal. People could at least apply this where possible and buy products where they're in control somehow

For storage platforms like Dropbox or Onedrive, it can be as simple as ctrl+c'ing the data out of there every now and then

For Spotify, you can do the GDPR export and check that you can open it somehow. Even if it's not super readable, you can figure that out (with a tech friend or LLM perhaps) if you turn out to ever need it

For Signal, I've got no idea. The format is hard to work with for hardcore techies. They really really don't want you, ahem, attackers to get that data in the clear. (Two-sided coin). Make sure to turn on the backups, write down the passphrase, and take them off your phone every now and then, and hope there's a restore method when you need it

Perhaps there's a market for this. An automated or semi-automated backup of many cloud services. It could probably be as simple as doing a gdpr request in some cases.

That's not necessarily simple though, because some companies are shady and ass about it. And then if you want to pursue the matter, be ready to face significant effort needed to file a proper complaint in the correct channels for that.

Fwiw, GDPR has a forgotten clause that data exports need to also be provided in machine-readable format for easy service migrations and promoting competition / reducing vendor lock-in

The only reason I can imagine why nobody implements this, is that they like the status quo where it's hard to switch to their service, but equally hard to switch away. Once this becomes commonplace (e.g. because vendor A makes users aware it's a thing by implementing it for migrating from vendor B), vendor B will "retaliate" and A will have to truly be better than vendor B to retain said users

That is all to say, this might yet totally be a viable and legally unblockable business avenue for a third party to initiate. I'm just not sure there's enough demand to warrant (read: pay for) the effort involved. Maybe if you can have an LLM read the api docs of the service where you want to do the import and autogenerate a mostly-working version?

If you give a shit about music you should not use Spotify at all, but that is probably a different discussion

It is, but I think most people would care to at least have the song list, even if they'd need to go out and buy the corresponding data if they stop using the streaming service, or manually enter it into another streaming service

Reply on news.ycombinator.com