Backing up across machines with low free space? (lemmy.nocturnal.garden)

submitted 2 days ago by tofu@lemmy.nocturnal.garden to c/selfhosting@slrpnk.net

17 comments fedilink hide all child comments

I run some services on machine A. I want to backup the data to machine B. It's around 40 gigabytes in thousands of small files. I would prefer to compress the data, but there's not enough space on machine A to keep the data twice (original + tar archive). I would like to avoid copying that many files via scp/rsync since it's taking forever and a bit fragile.

Any solutions on how to solve this? Is it possible to compress live on machine A and stream into the archive on machine B without the need to keep the big archive file on A?

Eventually, it's supposed to be automated and B has like the last 3 dailys, 1 weekly, 1 montly.

B can ssh into A.

top 17 comments

sorted by: hot top controversial new old

[-] thelittleblackbird@lemmy.world 2 points 1 day ago

Borgbackup is all you need, you even don't need to allocate the space in your A machine...

[-] tofu@lemmy.nocturnal.garden 1 points 1 day ago

Seems like Borg Backup Pull Mode does what I need, using sshfs:

https://borgbackup.readthedocs.io/en/stable/deployment/pull-backup.html

[-] thelittleblackbird@lemmy.world 1 points 12 hours ago

Certainly it is a good way to go, but in my experience try the usual work flow, it is way more stable and the cpu burden is share between both nodes -> faster backups

[-] tofu@lemmy.nocturnal.garden 1 points 11 hours ago* (last edited 11 hours ago)

Machine B is not publicly reachable though.

[-] thelittleblackbird@lemmy.world 1 points 11 hours ago

Then nothing to say, good luck with your setup :)

[-] tofu@lemmy.nocturnal.garden 1 points 10 hours ago

Thanks! :)

[-] qaz@lemmy.world 3 points 1 day ago

Perhaps Kopia? It supports compression and deduplication.

[-] tofu@lemmy.nocturnal.garden 2 points 1 day ago

Thanks, I'll have a look!

[-] fruitycoder@sh.itjust.works 2 points 1 day ago

Have the filesystem compress by default and ship that?

I know you can do it with zfs. Export a snapshot of a fs with compression enabled. I'm less sure if you can do it with a btrfs, lvm, or xfs though. I know they support fs level compression though.

[-] tofu@lemmy.nocturnal.garden 2 points 1 day ago

One machine is a cloud VM with LVM I think, the other is my NAS with zfs. I think with zfs to zfs it would easily work!

[-] redlemace@lemmy.world 6 points 2 days ago* (last edited 2 days ago)

To compress you need diskspace. Simple as that. Three daily from 40 Gb in small files? By thor, sure hope you think incrementals. Have you tried backupc?

But honestly, it's worth -and most reliable- to invest in a new bigger disk to solve the problem instead of working your way around it.

[-] tofu@lemmy.nocturnal.garden 1 points 2 days ago* (last edited 1 day ago)

It's a VPS and I don't want to increase the monthly cost - in fact I want to decrease it by disabling the backups the hoster offers. Machine B is my NAS at home.

Of course, incremental would be preferable, but I'm not sure how that would work in that constellation.

I wonder if I can compress to B directly. Maybe I could mount a volume from B into A via wireguard and nfs, now that I think about it.

[-] redlemace@lemmy.world 2 points 1 day ago

Do you have unlimited data transfer with your vps? Transferring 40Gb three times each and every day is quite some traffic.

Why not host it at home then? I bought a used lenovo thinkcenter (m720q on top of my head) for 75€ It uses only 35 watt, it has 750 Gb storage and I only added some RAM.

I pay 1€ per month for a fixed public ip to avoid ddns. My NAS is at a family member (over vpn to home) for offsite backups.

[-] tofu@lemmy.nocturnal.garden 2 points 1 day ago

I do have a someserver, but it's a fediverse instance with around 60 monthly active users, and the server provider is more reliable than my ISP at home, so any downtime has a bit more impact. There's also another admin who needs access. Not saying never, but at the moment I'm just looking for a different backup solution, moving the whole thing has some more aspects to consider.

The data would only be copied once a day. I have to check the limits though, good point.