I’m trying to plan a better backup solution for my home server. Right now I’m using Duplicati to back up my 3 external drives, but the backup is staying on-site and on the same kind of media as the original. So, what does your backup setup and workflow look like? Discs at a friend’s house? Cloud backup at a commercial provider? Magnetic tape in an underground bunker?
Currently only have pictures and documents stored, so everything easily fits on 1tb. One copy on my homeserver (unencrypted), one copy on my laptop (Luks encrypted), and one copy with rsync and a raspi at my parents (unencrypted). Might change encryption strategies to all luks.
I use Proxmox Backup Server for my backups. Everything backups to 1 system at home. I then sync the data store to a little NAS I have at a family members house across town and also to a cheap storage VPS on the other side of the country. I also do a manual sync of the data store to a single external drive that I manually connect and disconnect.
None of my data hoarding files are backed up as that would cost way too much. That could change if I ever find a killer deal on an LTO8 or better drive and tapes.
I know that Hetzner has some decently priced Storage Boxes that you can mount using rclone and then backup to. Keep in mind that latency will be a factor so it could be slow.
I don’t think this meets the definition of 3-2-1. Which isn’t a problem if it meets your requirements. Hell, I do something similar for my stuff. I have my primary NAS backed up to a secondary NAS. Both have BTRFS snapshots enabled, but the secondary has a longer retention period for snapshots. (One month vs one week). Then I have my secondary NAS mirrored to a NAS at my friends house for an offsite backup.
This is more of a 4-1-1 format.
But 3-2-1 is supposed to be:
Three total copies of the data. Snapshots don’t count here, but the live data does.
On two different types of media. I.e. one backup on HDD and another on optical media or tape.
I’ve always understood 2 as 2 physically different media - i.e., copies in different folders or partitions of the same disk is not enough to protect against failure of that disk, but a copy on a different disk does. Ideally 2 physically different systems, so failure/fire in the primary system won’t corrupt/damage the backup.
Used to be that HDDs were expensive and using them as backup media would have been economically crazy, so most systems evolved backup media to be slower and cheaper. The main thing is that having /home/user/critical, /home/user/critical-backup, and /home/user/critical-backup2 satisfies 3 copies, but not 2 media.
Hm I wonder why snapshots wouldn’t satisfy 3. Copies on the same disk like /file, /backup1/file, /backup2/file should satisfy 3. Why wouldn’t snapshots be equivalent if 3 doesn’t guard against filesystem or hardware failure? Just thinking and curious to see opinion.
If I’m reading your example right, I don’t think that would satisfy three either. Three copies of the data on the same filesystem or even the same system doesn’t satisfy the “three backups” rule. Because the only thing you’re really protecting against is maybe user error. I.e. accidental deletion or modification. You’re not protecting against filesystem corruption or system failure.
For a (little bit hyperbolic) example, if you put the system that has your live data on it through a wood chipper, could you use one of the other copies to recover your critical data? If yes, it counts. If no, it doesn’t.
Snapshots have the same issue, because at the root a snapshot is just an additional copy of the data. There’s additional automation, deduplication, and other features baked into the snapshot process but it’s basically just a fancy copy function.
Edit: all of the above is also why the saying “RAID is not a backup” holds true.
Right so I guess the question of 3 is whether it means 3 backups or 3 copies. If we take it literally - 3 copies, then it does protect from user error only. If 3 backups, it protects against hardware failure too.
E: Seagate calls them copies and explicitly says the implementer can choose how the copies are distributed across the 2 media. The woodchipper scenario would be handled by the 2 media requirement.
All storage is on a Ceph cluster with 2 or 3 disk/node replication. Files and databases are backed up using Velero and Barman to S3-compatible storage on the same cluster for versioning. Every night, those S3 buckets are synced and encrypted using rclone to a 10tb Hetzner Storage Box that keeps weekly snapshots.
Bit more than 3 copies, but hdd storage is cheap. Majority of my storage is Jellyfin anyways, which doesn’t get backed up.
I’m working on setting up some small nvme nodes for the ceph cluster, which will allow me to move my nextcloud from hdd storage into its own S3 bucket with 4+2 erasure coding (aka raid 6). That will make it much faster and also its cut raw storage usage from 4x to 1.5x usable capacity
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !selfhosted@lemmy.world
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don’t control.
Rules:
Be civil: we’re here to support and learn from one another. Insults won’t be tolerated. Flame wars are frowned upon.
No spam posting.
Posts have to be centered around self-hosting. There are other communities for discussing hardware or home computing. If it’s not obvious why your post topic revolves around selfhosting, please include details to make it clear.
Don’t duplicate the full text of your blog or github here. Just post the link for folks to click.
Submission headline should match the article title (don’t cherry-pick information from the title to fit your agenda).
No trolling.
Resources:
selfh.st Newsletter and index of selfhosted software and apps
“3! 2! 1!” Is just what I say when doing some potentially deleterious action after rsyncing a few key directories to a separate volume
3 sticky notes telling me to “go get that incremental backup working”,
2 separate external hard drives,
1 month out of date
Same lol. Can’t be that catastrophic. Right? …. Right?
borgmatic is way too easy and hetzner storage box is way too cheap to have any excuses
A usb stick and an old hard drive from 2009. The crackhead way of dealing with backups.
I dump my encrypted data to someone who probably practices 3-2-1 rule (which is Backblaze for me). I mean, these guys back up data for a living.
My current plan once new migration is completed:
Primary pool - 1x ZFS (couldn’t afford redundancy but no different to my RPI server). My goal is to get a few more drives and set up a RAIDZ1/2.
Weekly backup of critical data (eg. nextcloud) from primary pool to a secondary pool. Goal here is to get a mirror but will only be one drive for now.
Weekly upload of secondary pool to hetzner storage box via rsync.
Current server
1x backup to secondary drive (rpi) 1x backup to hetzner storage box via rsync
Currently only have pictures and documents stored, so everything easily fits on 1tb. One copy on my homeserver (unencrypted), one copy on my laptop (Luks encrypted), and one copy with rsync and a raspi at my parents (unencrypted). Might change encryption strategies to all luks.
I use Proxmox Backup Server for my backups. Everything backups to 1 system at home. I then sync the data store to a little NAS I have at a family members house across town and also to a cheap storage VPS on the other side of the country. I also do a manual sync of the data store to a single external drive that I manually connect and disconnect.
None of my data hoarding files are backed up as that would cost way too much. That could change if I ever find a killer deal on an LTO8 or better drive and tapes.
I know that Hetzner has some decently priced Storage Boxes that you can mount using rclone and then backup to. Keep in mind that latency will be a factor so it could be slow.
DO NOT follow my lead, my backup solution is scuffed at best.
3:
I have:
2:
I’ve got a hard drive and flash memory?
1:
Don’t have this at all, the closest is that my phone is off-site half of the day.
Real selfhosters know
my backup is staring longingly at LTO drives and wishing they would magically be affordable.
Does this make sense?
I don’t think this meets the definition of 3-2-1. Which isn’t a problem if it meets your requirements. Hell, I do something similar for my stuff. I have my primary NAS backed up to a secondary NAS. Both have BTRFS snapshots enabled, but the secondary has a longer retention period for snapshots. (One month vs one week). Then I have my secondary NAS mirrored to a NAS at my friends house for an offsite backup.
This is more of a 4-1-1 format.
But 3-2-1 is supposed to be:
Three total copies of the data. Snapshots don’t count here, but the live data does.
On two different types of media. I.e. one backup on HDD and another on optical media or tape.
With at least one backup stored off site.
I’ve always understood 2 as 2 physically different media - i.e., copies in different folders or partitions of the same disk is not enough to protect against failure of that disk, but a copy on a different disk does. Ideally 2 physically different systems, so failure/fire in the primary system won’t corrupt/damage the backup.
Used to be that HDDs were expensive and using them as backup media would have been economically crazy, so most systems evolved backup media to be slower and cheaper. The main thing is that having /home/user/critical, /home/user/critical-backup, and /home/user/critical-backup2 satisfies 3 copies, but not 2 media.
Hm I wonder why snapshots wouldn’t satisfy 3. Copies on the same disk like /file, /backup1/file, /backup2/file should satisfy 3. Why wouldn’t snapshots be equivalent if 3 doesn’t guard against filesystem or hardware failure? Just thinking and curious to see opinion.
If I’m reading your example right, I don’t think that would satisfy three either. Three copies of the data on the same filesystem or even the same system doesn’t satisfy the “three backups” rule. Because the only thing you’re really protecting against is maybe user error. I.e. accidental deletion or modification. You’re not protecting against filesystem corruption or system failure.
For a (little bit hyperbolic) example, if you put the system that has your live data on it through a wood chipper, could you use one of the other copies to recover your critical data? If yes, it counts. If no, it doesn’t.
Snapshots have the same issue, because at the root a snapshot is just an additional copy of the data. There’s additional automation, deduplication, and other features baked into the snapshot process but it’s basically just a fancy copy function.
Edit: all of the above is also why the saying “RAID is not a backup” holds true.
Right so I guess the question of 3 is whether it means 3 backups or 3 copies. If we take it literally - 3 copies, then it does protect from user error only. If 3 backups, it protects against hardware failure too.
E: Seagate calls them copies and explicitly says the implementer can choose how the copies are distributed across the 2 media. The woodchipper scenario would be handled by the 2 media requirement.
4-2-1-1 for me I guess 🫣 or 4-2-2?
Two copies at home, synced daily, one of them in an external drive that I like to refer as the emergency grab and run copy lol
One at a family member synced weekly and manually every time I visit.
All of those three copies are always within a 10 kilometer radius in a valley overseen by a volcano so…
One partial copy of the so-critical-would-cry-if-Iost data is synced every few days to a backblaze bucket.
I rawdog storage. I RAID0 and forget. huehue.
Toss in another drive for RAID5. That way you can at least have some redundancy…
It’s not important data. Why would I spend another $200+ for another 20TB drive to have redundancy for 1 and 0 I don’t care about…
Fair point.
I use Kopia to B2, then on a monthly basis I copy the current Kopia repo to an external drive that’s otherwise kept offline in my house.
All storage is on a Ceph cluster with 2 or 3 disk/node replication. Files and databases are backed up using Velero and Barman to S3-compatible storage on the same cluster for versioning. Every night, those S3 buckets are synced and encrypted using rclone to a 10tb Hetzner Storage Box that keeps weekly snapshots.
Config files in my git repo:
https://codeberg.org/jlh/h5b/src/branch/main/argo/external_applications/velero-helm.yaml
https://codeberg.org/jlh/h5b/src/branch/main/argo/custom_applications/bitwarden/database.yaml
https://codeberg.org/jlh/h5b/src/branch/main/argo/custom_applications/backups
https://codeberg.org/jlh/h5b/src/branch/main/argo/custom_applications/rook-ceph
Bit more than 3 copies, but hdd storage is cheap. Majority of my storage is Jellyfin anyways, which doesn’t get backed up.
I’m working on setting up some small nvme nodes for the ceph cluster, which will allow me to move my nextcloud from hdd storage into its own S3 bucket with 4+2 erasure coding (aka raid 6). That will make it much faster and also its cut raw storage usage from 4x to 1.5x usable capacity
Sometimes: a laughing hyena.
If you don’t have tested backups, you don’t have a backup.