I have a 56 TB local Unraid NAS that is parity protected against single drive failure, and while I think a single drive failing and being parity recovered covers data loss 95% of the time, I’m always concerned about two drives failing or a site-/system-wide disaster that takes out the whole NAS.
For other larger local hosters who are smarter and more prepared, what do you do? Do you sync it off site? How do you deal with cost and bandwidth needs if so? What other backup strategies do you use?
(Sorry if this standard scenario has been discussed - searching didn’t turn up anything.)
I’m not sure if I qualify as a ‘larger local hoster’ but I would go through your 20 TB and decide what really is important enough to backup in case the wheels fall off. Linux ISOs, those can be re-downloaded, although it would take a bit of time. The things that can’t be readily downloaded such as my music collection that I have been accumulating for decades, converted to flac, and meticulously tagged, can’t be re-downloaded. So that is one of my priorities to back up. Pictures, business documents, personal documents, can’t be re-downloaded, so that goes on the ‘must back up’ list…and so on. Just cull out what is and isn’t replaceable. I would bet that once you do that, your 20 TB will be a bit more slim, and you’re not trying to push 20TB up the pipe to a cloud backup.
I use BackBlaze’s Personal, unlimited tier for $99 USD per year, which is a pretty sweet deal. One thing about Backblaze to remember is that the drives being backed up must be physically connected to the PC doing the backup/uploading. I get around that because I have a hot swap bay on my main PC, but there are other methods and software that will masquerade your NAS or other as a physically connected drive.
I use backblaze too, started with the personal back up, but swapped to the B2 solution as it was supported by my NAS. The cost of the actual storage isn’t much, most of the cost is in access, so for data that doesn’t alter much it worked out just as cheap, and easier to do things that way.
Lto tape. But I only have 15tb
It quickly becomes cost effective when you actually need the data to be safe. Far easier to have off site backups. I have never had a problem , but I like to have offline backup. Most of the time my data is static. So I am only backing up projects files ans changes for the most part.
If you have 40+ tb of dynamic data I can’t help there.
As someone who has experienced double failure twice in my lifetime, I seriously recommend doing backups.
The problem is that the only serious backup solution is another HDD for this size. A robot array for tapes or worm drives is probably out of budget.
I don’t. Of my 120tb, I only care about the 4tb of personal data and I push that to a cloud backup. The rest can just be downloaded again.
Do you have logs or software that keeps track of what you need to redownload? A big stress for me with that method is remembering or keeping track of what is lost when I and software can’t even see the filesystem anymore.
I don’t know of a pre-wrapped utility to do that, but assuming that this is a Linux system, here’s a simple bash script that’d do it.
#!/bin/bash # Set this. Path to a new, not-yet-existing directory that will retain a copy of a list # of your files. You probably don't actually want this in /tmp, or # it'll be wiped on reboot. file_list_location=/tmp/storage-history # Set this. Path to location with files that you want to monitor. path_to_monitor=path-to-monitor # If the file list location doesn't yet exist, create it. if [[ ! -d "$file_list_location" ]]; then mkdir "$file_list_location" git -C "$file_list_location" init fi # in case someone's checked out things at a different time git -C "$file_list_location" checkout master find "$path_to_monitor"|sort>"$file_list_location/files.txt" git -C "$file_list_location" add "$file_list_location/files.txt" git -C "$file_list_location" commit -m "Updated file list for $(date)"That’ll drop a text file at
/tmp/storage-history/files.txtwith a list of the files at that location, and create a git repo at/tmp/storage-historythat will contain a history of that file.When your drive array kerplodes or something, your files.txt file will probably become empty if the mount goes away, but you’ll have a git repository containing a full history of your list of files, so you can go back to a list of the files there as they existed at any historical date.
Run that script nightly out of your crontab or something (
$ crontab -eto edit your crontab).As the script says, you need to choose a file_list_location (not
/tmp, since that’ll be wiped on reboot), and set path_to_monitor to wherever the tree of files is that you want to keep track of (like,/mnt/file_arrayor whatever).You could save a bit of space by adding a line at the end to remove the current files.txt after generating the current git commit if you want. The next run will just regenerate
files.txtanyway, and you can just usegitto regenerate a copy of the file at for any historical day you want. If you’re not familiar withgit,git logto find the hashref for a given day,$ git checkout <hashref>to move where things were on that day.EDIT: Moved the
git checkoutup.That’s incredibly helpful and informative, a great read. Thanks so much!
My *arrstack DBs are part of my backed up portion, so they’ll remember what I have downloaded in my non-backed up portion.
That’s a great point.
In my case, for Linux ISOs, is only needed to login in my usual private trackers and re-download my leeched torrents. For more niche content, like old school TV shows in local language, I would rely in the community. For even more niche content, like tankoubons only available at the time on DD services, I have a specific job but also relying in the same back up provider that I’m using for personal data.
Also, as it’s important to remind to everyone, you must encrypt your backup no matter where you store it.
If you can’t remember what you lost, did you really need it to begin with?
Unless it’s personal memories of course.
I can’t remember the name of an excel spreadsheet I created years ago, which has continually matured with lots of changes. I often have to search for it of the many I have for different purposes.
Trusting your memory is a naive, amateur approach.
You put that with everything else similar into a folder, which is backed up. Mine is called “Files”. If there’s something in there that I don’t need backed up. It still gets backed up. If there’s something very large in there that I don’t need backed up, it gets removed in one of my “oh shit these backups are huge” purges.
The key here being that you actually remember the file exists, because it’s important. Some other random spreadsheet you don’t even remember exists because you haven’t needed it since forever is probably not all that important to backup.
If you loose something without ever realizing you lost it, it was not important so there would be no reason to make a backup.
So you do remember that you have several frequently-used spreadsheets.
For me, I have a bad memory. I might remember a childhood movie (a nickname I give to special Linux ISOs) that I hadn’t even thought of for 10 years and track down a copy, sometimes excavating obscure sources, and that may be hours of one-time inspiration and work repeated many times over. Having a complete list is a good helper, but a full backup of course is best.
servarr* and jellyfin are managing my movies and tv-shows
Set up a job to write the file names of everything in your file system to a text file and make sure that text file gets backed up. I did that on my Unraid server for years in lieu of fully backing up the whole array.
That should be part of the backup configuration. You select in the backup tool of choice what you backup. When you poose your array then you download that stuff again?
I only care about the 4tb of personal data and I push that to a cloud backup
I have doubles of the data. Some of 'em. That way I know I have a pristine one in backup. Then I can use it, it gets corrupted, I don’t care.
Actually, I have triples of the W2s. I have triples, right? If I don’t, the other stuff’s not true.
See, the W2s the one I have triples of. Oh, no, actually, I also have triples of the kids photos, too. But just those two. And your dad and I are the same age, and I’m rich and I have triples of the W2s and the kids photos.
Triples makes it safe.
Triples is best.
Bob Odenkirk has never steered us wrong, thanks. I downloaded three copies of this from YouTube in case I forget.
Same here, ~30TB currently but my personal artifacts portion is only like 2TB, which is very affordable with rsync.net, which conveniently has an alerts setting if more than X kb hasn’t changed in Y days. (I have my Synology set up to spit out daily security reports to meet that amount, so even if I don’t change anything myself I won’t get bugged)
I follow a similar strategy. I back up my important stuff. And I’m gonna have to re-rip my DVD collection and redownload the Linux ISOs in the unlikely case the RAID falls apart. That massively cuts down on the amount of storage needed.
Backup to 2nd nas.
Important stuff gets backed up to cloud storage. Whatever is cheapest.
In my case Synology c2 cloud was cheapest.
Recently helped someone get set up with backblaze B2 using Kopia, which turned out fairly affordable. It compresses and de-duplicates leading to very little storage use, and it encrypts so that Backblaze can’t read the data.
Kopia connects to it directly. To restore, you just install Kopia again and enter the same connection credentials to access the backup repository.
My personal solution is a second NAS off-site, which periodically wakes up and connects to mine via VPN, during that window Kopia is set to update my backups.
Kopia figures out what parts of the filesystem has changed very quickly, and only those changes are transferred over during each update.
The Backblaze option is something I’ve seriously considered.
Any reason this person didn’t go with the $99/year personal backup plan? It says “unlimited” and it is for my household only, but maybe I’m missing something about how difficult it is to setup on Unraid or other NAS software. B2’s $6/TB/mo rate would put me at $150/mo which is not great.
You can’t use the $99/year plan for that. The authorized client only works as a desktop application on Windows and MacOS.
They only needed about 500GB.
And personal is for desktop systems. You have to use Backblazes macOS/Windows desktop application, and the setup is not zero-knowledge on Backblazes part. They literally advertise being able to ship you your files on a physical device if need be.
Which some people are ok with, but not what most of us would want.
Like others, I have a 2 tier system.
About 2TB of my (Synology) NAS is critical files. Those get sent via Hyperbackup to cloud storage on at least a weekly basis, some daily. I have them broken up into multiple tasks with staggered schedules so it never has much to do on any given day.
The other 16TB I have get sync’d (again with hyperbackup, but not a scheduled backup task) to a 20TB external drive roughly once per quarter. Then that drive lives on the closet of a family member.
entire nas (~24TB used) is replicated to another nas in another building (2 actually). i like having 3 copies.
Similar to most responses, I backup whatever I created myself, not shared by someone or downloaded from somewhere. I care about pictures that I took, documents, financial records, etc, which don’t take up much space at all.
Acronyms, initialisms, abbreviations, contractions, and other phrases which expand to something larger, that I’ve seen in this thread:
Fewer Letters More Letters Git Popular version control system, primarily for code NAS Network-Attached Storage RAID Redundant Array of Independent Disks for mass storage SSD Solid State Drive mass storage VPN Virtual Private Network
5 acronyms in this thread; the most compressed thread commented on today has 7 acronyms.
[Thread #119 for this comm, first seen 26th Feb 2026, 15:51] [FAQ] [Full list] [Contact] [Source code]
I have 2 500GB SSDs in RAID1 for important data, truenas apps etc…, then 32TB total in RAIDZ1 for large Dataset that won’t need speed (movies, TV show, music, pictures, archives,…)
If I have a complete NAS failure, a remote backup (via rsync to a friend’s NAS Weekly) of the SSD and bootable drive can be used in a new system, and my torrent app has the list and magnet of all torrents stored on the SSD so it can re-download them.
The stuff that I actually care about are automatically backed up twice, once to a simple external on site and once to a cloud. The cloud rotates between the most recent backups so it never takes up more than 1tb compressed, while the local external keeps backups for much longer (something like 6tb at a time).
So you have 56TB of total storage, but how much of that 56TB is actually used? Take the amount of storage used and add 10-12% to that figure. Now you create a new NAS (preferably off-site) with that amount of storage and that becomes your backup target. Take an initial backup (locally if possible to speed up the process) and then you can use something like rsync to create incremental backups going forward. This is the method I’ve used and so far it has worked out well. I target 10-12% more than the amount of used storage for my backup capacity because my storage use grows reasonably slowly. If your usage grows faster you might want to increase your “buffer” a little more so that you’re not having to constantly add drives to your backup target.
Yeah, this is certainly a viable “brute-force”-ish ooption. While I have 56, I’m only using 26 or so. But I’d actually be hesitant to do anything less than a full capacity mirror because I do expect to eventually use this (and more - adding drives to Unraid).
I’ve balked because of cost and upkeep (maintaining the same capacity, additional chances for drive failure, two separate sites I need physical access to with a high bandwidth connection), so I admit I was hoping I was missing an easier option.
I mean, if you want a full mirror, rolling your own backup target is going to be the cheapest option even with the current high price of hardware. Other options are cloud storage, or using another media like tape. Cloud storage is of course an on going cost which rules it out for me, not to mention privacy concerns. There are certain “cold storage” options from cloud storage hosts which are considerably cheaper but they have limitations on how the data can be accessed and how often. The tape route is possible but it’s not really viable for home use due to the high upfront cost of the drives. Outside of that, backing up a subset of your storage as others have suggested is the only other option. Creating viable backups without breaking the bank is a challenge as old as computers, unfortunately.
