Hi, I’ve been using CloudBerry since '15 successfully to Glacier. Has been on an older XP machine (!) which pulls from a NAS and pumps the data up to Amazon. This is a kludgey system. So need to re-do it.
BIG QUESTION: What’s a best backup recipe that will enable file versioning (like Time Machine)? And which has BOTH a local NAS backup server AND a cloud backup?
CONCERN: We had a nice NAS device on SMB LAN networking. And we could put CloudBerry directly on the NAS. However, with the threat of ransomware, there’s the danger that a problem on any member of the LAN desktops could propagate to the shared files NAS.
SOLUTION: Two NASes: 1 x sharing and 1 x backups
(1) FILESHARING NAS – for file sharing and large files, serving everyone on the LAN.
(2) BACKUP NAS – The filesharing NAS rsyncs over to a second NAS that is not on SMB LAN networking. The connection would be via SFTP (i.e. SSH).
This seems to be a recommended solution when you get beyond the basics. Our sense is that basic backup is mostly too basic.
CURRENT LAN CONFIGURATION: W10 desktop clients, shared BSD NAS device, all over wired Ethernet and also some WiFi for laptops.
PROPOSED SOLUTION:
-
DESKTOP BACKUPS: Considering doing rsync (or if there’s a W10 GUI version of this) or other backup software to the BACKUP NAS over SFTP. Full near real-time backups but no risk. Important that VSS is working.
-
FILESHARING NAS: Same as desktop backups, SFTP to BACKUP NAS, with no risk.
-
BACKUP NAS TO CLOUD: Put CloudBerry (re-purposed license to Linux) on Linux in a VirtualBox Guest on the BSD NAS. Use CloudBerry to pump versioned files to some cloud host. 2.5 TB total without duplication for versioning.
BEST PRACTICES QUESTIONS
a) VERSIONING - How can we do file versioning, rather like TIme Machine on Mac? This is essential for protection against ransomware.
b) THRASHING With four or five clients pumping files into the NAS Backup NAS, plus Cloudberry pumping files up the Cloud, will this create disk thrashing on the Backupu NAS? Or is the server smart enough to queue things, and it just takes longer? Probably there won’t be that many files eligible for backup in any given hour or two anyway.
c) DIRECTORY CHANGE PROPAGATION – If a directory name change takes place, or even if a directory is moved, this can create cascading unnecessary changes. How does CloudBerry deal with this? And similarly, how would Cloudberry deal with deletions, as the propagate from desktop to Backup NAS to cloud?
d) ARCHITECTURE – Is the above plan the best architecture? Let’s just ignore the file sharing NAS, and think of this recipe as A) first, backup local clients to local backup NAS and then (B) backup local backup NAS to remote Cloud.
SUMMARY RECIPE
I’m hoping that someone here will have some professional insight here. Getting this right could be helpful to a lot of people.
Thanks,
JM