I’m brand new to UrBackup, and generally ignorant of all but the simplest backup platforms, but hopeful.
Our organization has a (very sensible) requirement to incorporate offline, offsite copies in our backup strategy for about 150TB across 20 linux servers. Much of this data is static and can be identified, tarred+gzipped to a mess of HDDs in a file cabinet. But a significant amount, say 30TB (and growing), requires some kind of scheme for rotating media offline on a tolerable loss frequency (say weekly). We’d also like to do this without incurring cloud costs since our LAN is free and we have the physical HDDs on hand.
I imagine it would be easy if I could tell UrBackup to stop after the first full backups are done, then remove my storage backend (say a few drives RAID’d or LV’d together) while saying “that’s full backup v1”, and then tell UrBackup “we’re good, ignore that you no longer find v1 files on the storage, just go ahead and write differentials from v1”, and forever rotate differential-only drives.
Is this something that can be done with UrBackup, e.g. something like the old tape backup systems?
Apologies if I’m not using appropriate jargon (or seeing an obvious solution in the docs or settings since I don’t know the jargon).
Can t really find a not contorted way to tick all your cases with urbackup. You would probably need to use btrfs send/receive. My issues stems from the fact that you would need to archive raid drives, so manipulate many disks. And those would not be available in urbackup.
If i were you, i would try to compromise, with a second, remote archival server for long term retention.
Tapes use the same logic that you are, and it’s very error prone, with missing tapes that cause unusable backup. Or corrupt ones if you have them all so you need to periodically re-read them, so use a big robotic, which expensive, send alerts, so it is online blah blah blah, it’s a pain for nothing.
You may want to have a look at dar backup. http://dar.linux.free.fr/doc/index.html
Its originally made to archive cd, while providing an online catalog.
It’s a pita to understand, it’s command line options are complex, and it has no gui.
But it basically does the exact same thing you are doing right now.
If you can write scripts to automate it, it may be good for you.
Thanks, orogor. Doesn’t necessarily need to be raids, but regardless it would indeed be a lot of disks offline to manage. I’ve been hoping to find a consolidated approach to our backups, but think in the end I will have to cobble together multiple solutions, with UrBackup possibly handling the smaller, frequent backups.
It would be great to see in the documentation a few diverse use cases laid out to understand the pros/cons/applications of UrBackup. It has so many options and features (amazing!) that it can be hard for a non-expert to know where to begin without quite a bit of trial and error. edit: just found this awesome bit: https://www.urbackup.org/administration_manual.html#x1-10800011.7 more like this, please!
For example, putting aside offline backup, what do people do when an increase in # or size of clients exceeds the backup storage? Add a 2nd backup device and run a 2nd UrBackup server, and shift some clients to that?
You can’t add a second datastore to urbackup, so either you can resize your storage (san) , or you use a different server. On the free version there’s no federated storage, to help with multiple servers, but it s present in some paying versions (aws/virtual appliance, can t remember which) which aren’t expensive.
If you want to use image backup , i stringly suggest you get a single cbt license to try out 15€, and set your storage to btrfs+raw images