I currently have a 6 terabyte aligned in a Raid1 configuration to run my 3 terabyte cloud server at home.
I live in a small studio apartment but I realized that I had a spare power supply, motherboard, and graphics card in my closet and so it seems like it wouldn’t really be that big a deal for me to just buy another case and split my server up into two servers.
one advantage of this is that I currently have my server set up in a bit of a strange configuration so as to have access to certain files. I have sync client installed on the server so that a select number of files are available outside of my www/nextcloud folder though this can cause problems with syncing if I manipulate a file and then don’t let it sync before I mess with it.
If I divided my servers off into separate boxes, then I could have one of the boxes act as just a media server, using the sync client, and the other could just be a nice normal server with no fancy weird redundant configurations like I currently have.
The other advantage I see in doing this is that I could physically separate the HDDs more for added redundancy. A RAID1 server is redundant but if something like an electric shock or something happened it could potentially fry the entire thing, right? Having my media server and home server both in the same building isn’t quite as safe as having offsite backup or whatever but it seems better than having them in the same box, right?
Offsite backup is always a very good idea, even if data is not the newest. In case of fire, burglary etc. you have something even if it’s a month old.
Offline backup is a very good thing as well, if your system crashes or files are deleted by accident (or malfunction).
So do you just want to avoid data loss or is it important, that you can get up a broken system quickly (10 min, 1 h, 1 day or 1 week)? A clean setup has always the advantage that it is easier to backup&restore, as well as running upgrades. I tend to keep a clean productive setup and use a second for testing and fun stuff where things can fail. Running two servers on the other hand uses more power. If they are big enough, virtual machines within the same server are a good option as well (it’s then as well very easy to move a virtual machine to a different host machine).
By power you mean like extra electricity right? I don’t think that should be too much of an issue. A second computer is like a couple extra bucks a month in electricity after all.
I’m also not too worried about quick recovery more just about making sure data is as safe as reasonably possible and without me having to spend extra money on like Amazon servers or something.
So you think having a second computer using the sync client is sufficient as an additional back up? Or is there some additional protocol I should be using and why?
first you should define want what you need/want as clearly as possible. then find the technically (almost) perfect solution for this and find out how close you can get to it with the means at your disposal - find a good and working compromise. what you describe sounds just the other way around. (but of course gaining experience by a hardware-experiment has value, too.)
“redundancy” (whatever kind you mean by it) does usually not have a high priority in private data setups.
you’re probably much better served if you get some external hd-cases for your excess hard-drives (get rid of raid-1) and put backups on them.
completely backup your existing server, plan its disk-layout (i like btrfs bc of stuff like subvols, snapshots etc) and re-install it in a simple configuration (no raid). restore the backup.
after that, make regular backups on the harddisks, store them at different places (girlfriend, parents, work, …)
it can be advantageous to use different backup-programs: filesystem-based stuff like rsync, rsnapshot or btrfs-tools tends to be slower (and sometimes more space-consuming) in backup but faster in restore (especially for whole systems) while more sophisticated stuff (like borg, duplicity, obnam restic, …) is faster (and (sometimes considerably) less space-consuming) in backup but of course requires you to have the software installed (and often access to encryption-keys etc) in the case of a restore.
(as a non-technical aside i would like to mention that thinking about electricity exclusively in terms of personal means ($$$) precludes the possibility of (m)any future generations inhabiting this planet in a way similar to ours).
GOOD LUCK!
(as a non-technical aside i would like to mention that thinking about electricity exclusively in terms of personal means ($$$) precludes the possibility of (m)any future generations inhabiting this planet in a way similar to ours).
Agreed, it’s just that the amount of electricity used by a home desktop is dwarfed by industrial usage of electricity (something like 100 companies account for 70% of the global carbon emissions), but fair enough, point well taken.
For now I’ve decided to keep things the way they are and I will look into making a second server and finalizing some kind of standardized offline backup later in the future. I’ve got a lot of other stuff I need to handle first