A quick answer to your immediate question: which Ubuntu version to use? The one that empowers you to spend more time on photography and less time worrying about keeping the system running. But that answer raises a lot of follow up questions. How deep do you want to dive into Linux?
At the risk of expressing opinions not facts, here are a few pointers from my own experience, in order of importance.
Bare Metal (parts list):
Redundancy (RAID Level):
- What is the most important factor? capacity (total vs.usable)? performance? data preservation? ease of recovery? peace of mind? rank your factors
- Fact: Simplicity makes recovery easier. The higher the RAID level, the higher the complexity.
- Opinion: After 20 years of experience with different levels of hardware RAID and software RAID, I swear by mirroring only. If I need more capacity, I add another mirror. If I need more redundancy, I add another drive to the mirror. I value the integrity of my data and the economy of my time over the optimization of capacity and performance. I currently use 6 drives plus cloud (see cloud section below). 3 drives in the off-site backup (two are ZFS mirrors and one is spare). Two in the local server (ZFS mirror). One for offline backup (see below)
Scale
If the current generation of drives is not large enough to fit your data, do not bother with technical solution: just set up a second or third RAID. I did it by grouping years onto individual RAIDs. Revisit at every cycle of hard drives replacement. There is a golden point capacity at every given moment in history (price/GB). Choose what works for you when you need to replace your current drives and when migrating the data from the previous generation to the current one, decide how to split it. As I moved into another line of business and the quantity of photos / videos I produced diminished, disk capacity caught up with my storage needs.
File System
Opinion: I am using ZFS since 2008. The reason is personal circumstances. My archive dates back to the days of Commodore Amiga and has been on all sorts of file systems. At the turn of the previous millenium, I moved it from NTFS on hardware RAID to ext2. It even had a stint on ReiserFS. I was using Red Hat Linux when Red Hat pulled the plug too hard and I found a new home with FreeBSD 5. FreeBSD introduced ZFS and from that day on I never used hardware RAID again.
Fact: There is healthy competition for modern file systems out there. btrfs, Ceph, ZFS are all great and further progress is being made. Ubuntu has fully adopted ZFS and as such it is a choice of convenience (I currently run a homelab with a few LXD containers and kvm virtual machines).
Backup Software
There are a lot of powerful tools out there. Depending how deep you want to get into the Linux ecosystem and learning new tools, my personal opinion is: learn about rsync and cron jobs.
There is an initial learning curve, but it is worth the effort. I have scripted rsync to:
automatically backup time-machine style my workstation to a 1TB SSD that I keep in my pocket with my current work https://www.amazon.com/dp/B0788CQYDN
rsync --link-dest=/path/to/previous-time-capsule /home/me/work /mountpoint/backup/current-time-capsule
regularly backup my local storage server to my off-site backup using ssh key logins (no passwords)
rsync /tank1 me@remote.server.com:/tank1
backup my local storage server to an offline disk when inserted https://www.amazon.com/dp/B000KS8S9W . This requires some udev rules (question: how deep into Linux do you want to dive? there are big benefits deeper down)
backup to/from the cloud. I run two instances of https://nextcloud.com/ -- one in the cloud and one in my homelab. Our family cellphones are set up to automatically beam all photos/videos to NextCloud.
The key piece of advice is to invest your time in something that is working for you, can grow with your needs, and will not drag you down a road you do not want to travel. I found that investing in low level knowledge has paid over the year, as rsync, ssh, udev, python have all evolved gently, nicely, respectful of not wasting existing knowledge of their user base and in a way to accommodate the many opinions and uses, including mine, These low-level tools have enabled me to achieve my objectives better than reliance on higher-level tools that are often predicated on the opinion of their owners. Their way or the highway. If you are comfortable building your own hardware, you can be comfortable with building-block software rather than shrink-wrapped solutions, and reap comparable benefits.
Offline Backup
No matter what choice of drives you make, all of the above so far is online. Connected to the internet and the dangers of ransomware encryption. Connected to a powerline. I always recommend to keep at least one off-line copy as well. It does not get updated immediately, and there is manual work involved which means the frequency of the updates depends on human factors. But it is an essential component. Keep at least one offline backup. Ideally one close by and one in a safety box somewhere not too far away (my off-site backup is a one day car drive away, YMMV depending on where your relatives live).
Cloud
Last but not least, the cloud, both as backup and as place to be backed up. If you do not publish your pics yet, you will at some point. In my opinion it makes sense to invest in a Backblaze, GoogleDrive, DropBox, and use them as additional backup. Some data is priceless. These services have in my opinion a place in every comprehensive backup plan. In my case, I have unlimited GoogleDrive space from my university alumni program and I encrypt and upload to the GoogleDrive pictures. I would probably use Backblaze for their software if GoogleDrive was not free. The important thing is to consider cloud storage like any D in RAID: it can and will fail. Design your system around it.
