I have about 500GB of data (photos, documents, videos etc.) that I have accumulated over the years. Currently, I keep them on my computer and rsync all additions / changes once a month or so to an external hard drive. Do I need to be worried about data loss (sectors going bad, bit rot, bit flip, whatever it is called)?
To clarify,
-
None of this is commercially important; I just don’t want to get into a situation where I look up an old family photo or video twenty years down the line and it has got corrupted.
-
Both my computer and the external HD are HDDs. They are fairly cheap here (and very cheap if second hand). Buying SSDs or dedicated hardware would be expensive.
The 3 2 1 rule is always the gold standard.
I’d recommend at least adding an offsite backup. Set up rclone with a mounted folder (client side encryption is recommended) and sync the files to that as well.
I use Backblaze for about $6/TB/mo, pro-rated for whatever amount is actually used.
second, for the small amount a backblaze account would be cheap and more than enough. If OP is worried about security then enabling a crypt endpoint in rclone is moderately trivial.
3-2-1 OP. 3 copies of your data, across 2 different storage mediums, with at least 1 offsite.
6$ is about 500 rupees. I can get another HDD for double that price.
I do copy some important files to Google Drive, but I don’t pay for it, and I don’t rely on it.
If you don’t pay for it, you can’t rely on it
Right, which is why I prefer to rely on local backups. Much cheaper in the long run.
I used to work with a guy who was religious about backing up his files to an external drives. Until someone broke into his house and stole his computer AND his external drives. He lost everything.
It’s always a good idea to have an off-site backup (e.g. in case of fires, robbery, natural disasters, etc). If you prefer to manage them yourself, an option is to find someone else who also needs an off-site backup and exchange disk space. You do your off-site on their machine, and they do theirs on yours. With external HDDs, you can just have someone else hold on to it for you at a different location. You can come up with fancier schemes to reduce the chances of data loss or to make the process simpler if you care to do so.
In my experience, a well treated, non overused physical hard drive can and likely will hold up for over 15 years.
I haven’t had any problems with any of my HDDs, but I don’t stress them out with daily gaming or video production, and I don’t toss them around like footballs, obviously.
Just speaking from my own experiences though…
My external HD is working well, but the computer’s HD seems to be of poor quality. I’m worried that once the primary copy gets corrupted, the mistakes will then be copied to the external HD as well. (Although if I understand rsync correctly, this shouldn’t happen.)
I started using restic for backups.
Pro:
- Encryption
- Deduplication
- Flexible backup location
- Data integrity checks
Con:
- No good GUI
Hard drives can fail. A strong magnetic field could scramble the data on the platters. HDD’s are pretty reliable usually though. Biggest concern with external HDDs would be fall damage.
I would say to check random files from time to time and you should be fine. Every 2 or 3 years, replace your backup drive. A backup program like Borg could help detect if you have a problem with your files, but you lose a bit of the simplicity of your current rsync method.
Anything your truly worried about should follow the 3,2,1 standard. Minimum 3 copies, on 2 separate media types, with 1 copy offsite. That said your current setup is already better than 95% of the general population and probably 70% of the Fedi.
I also just do this. However I have already found 2 photos that got randomly corrupted, and I don’t know how to prevent that.
So far my only idea was using md5sum, but checking all files like that takes a loooooooooong time.
I am paranoid about cloud. I do have my music backed up on OneDrive, encrypted with GPG using AES256, but I don’t even fully trust that. I know, it sounds stupid, but maybe in the future it will be quite easy to break.
But I don’t know much about encryption. Just reading the man page, I put these options together:
--s2k-cipher-algo AES256 --s2k-digest-algo SHA512 --s2k-mode 3 --s2k-count 65011712but whether I can consider that safe enough, I don’t know.
And since I don’t know enough about it, I prefer not to trust it.
I also just do this. However I have already found 2 photos that got randomly corrupted, and I don’t know how to prevent that.
If you are fine with changing your file system, check out zfs. It stores checksums with your data, and can, if configured to store multiple copies, repair corruption.
Instead of a single external HD set up a NAS with a raid configuration so that even if a drive fails the data is safe.
I want to point out that RAID doesn’t actively prevent bit rot and data degradation. You’ll want ZFS/RAIDZ for that.
I’m not super familiar with ZFS/RAIDZ, I guess it does extra data scrubbing and stuff to prevent data issues?
That’s cool. But a traditonal RAID setup still gives you redundancy and fault tolerance which is the important part, right?
It’s a software RAID integrated in the filesystem, as I understand it. This video helped me understand it a bit more and it’s why I’m saying ZFS is a better idea. afaik you get the good parts of raid and some more. Obviously I have very superficial knowledge on all of this though, so I recommend doing your own reading :P
Here is an alternative Piped link(s):
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source; check me out at GitHub.
I recommend encrypting them locally and backup the encrypted copy on a cloud drive.
