Hi, I was lucky and got a bunch of very good SSDs (Intel NVME enterprise U.2) for free. I built some of them into my server but still have many left and plan to use them for backups. That means that I’ll only connect them occasionally and store them in a drawer or even off-site. I’m aware that SSDs are not meant for long term “cold” storage as they can lose data when not powered on. However, I still really want to use them. This brings me to my question: What do I need to do to prevent data loss? Is it enough to connect them to power every few months and let the controller do its thing? Do I have to mount them or even read/write to them? Is there anything else I have to consider?
Thank you a lot in advance.
You’d probably want to use something that can do verification and correction. I’m sure that software exists.
Personally, I’d sell them and buy something more suitable.
They should be powered on if you want to retain data on them long-term. The controller should automatically check physical integrity and disable bad sections as needed.
I’m not sure if just connecting them to power would be enough for the controller to run error correction, or if they need to be connected to a computer. That might be model specific.
What server OS are you using? Are you already using some SSDs for cache drives?
Any backup is better than no backup, but SSDs are really not a good choice for long-term cold storage. You’ll probably get tired of manually plugging them in to check integrity and update the backups pretty fast.
Jeez, what a poor use of data center SSD’s. I’d install them in a server and use them. Note that Intel has various grades even of “enterprise” SSD’s, with the ones intended for low write load using QLC and likely not having such great data retention either.
For backups, ideally you want a tape drive. Failing that, if the data volume is not too large, optical media like DVD-R or maybe BD-R. Otherwise, I don’t have a good suggestion other than online storage that is kept spinning and maintained. Don’t use spun-down hard drives without considerable redundancy. They are horribly unreliable.
My own stuff (encrypted using Borg Backup) is in Hetzner Storage Box which is around 11 euro/month for 5TB. Every so often I get an email saying that the box will be unavailable on such-and-such date/time due to maintenance. I assume that means a drive in the RAID array has failed and they are replacing and rebuilding it.
Burned optical media shelf life can be as little as 5 years, so I don’t think it should be recommended for long-term storage.
It hurts me too, to have no better use for them. I just have no hardware with enough PCI lanes to connect a significant number of them.
I should have mentioned that I have various other backups (including Hetzner Storage Boxes and continously spinning enterprise HDDs).
I just wanted to use them to store “another copy” because I have no other idea how to use them.
TBH you should probably sell them in that situation. How many are there? It is a nice problem to have, I guess. You could of course stand up another server and install them in that. Will they fit in raspberry pi 5’s?