I have one main workstation (running Fedora right now) and now I build myself a second machine that I want to use manly for backup purposes and to play around with VMs. This is the first time for me to try something like this and in my vision it would go something like this:
From my main machine I send a command over the local network, the second machine turns on, dose the backup and turns itself off again when finished.
I want to set up and play with a VM? Again, turn on the second machine over the network, set the VM up and run it on the second machine while controlling it from the main one.
Got a few drives lying around? Add them to the second machine to increase the total storage/redundancy.
Need files while not being home? Spin the second machine up remotely and grab them. This is not needed right now, but would be appreciated to be possible to add later.
A friend/relative wants to have an offsite backup? Add them to the backup functionality of my second machine. This is also not needed right now, but would be appreciated to be possible to add later.
The machine I build for this propose got the following specs:
CPU: Ryzen 7 Pro 4750G
RAM: 64 GB
SSD: 2 * 1TB (For OS/Hypervisor and VMs. I think I would like to set them up in RAID 1)
HDD: 4 * 4TB (For Backups. Thinking about RAID5, but what would you recommend?)
As a Noob in this field I need help to decide how and with what software to approach this to hopefully meet the criteria mentioned above. Or you could also tell me that these criteria are bullshit and why. I’m open to suggestions.
And since I’m not familiar with all the software in this field, I’ll later on definitely need a bit of help to set it up as soon as I decided on something. But that will probably be best in another thread, right?
Okay, I understand the concept and am confident to be able to get it working after a bit of time of trying and reading. Thanks @jode .
The big undecided point for me is just still what OS/Hyperviser to use. I watched some of the videos Wendell did on backups, hypervisors, ZFS, etc. and also understand what he is conveying, but since I never before used any of these software’s it is hard for me to pick one.
At least the backup portion is something that I would like to set up once and forget it. At best get like a disk health report at the end of every backup too.
So, I guess what I’m basically want is a recommendation, considering the five points I made in the first post, what software best fits these criteria?
Build in backup function and scheduler, some NAS functionality, plays nice with WoL, is able to host VMs and lastly lets me add or remove not matching drives to or from pools somewhat easily.
But considering my track record with anything software setup, I’ll still just end up wiping everything a couple of times and then half a year later about 80% of what I wanted is finally working as it should with no clue how to get to 100%.
You have the right attitude to get things working. Expect some effort to get things working initially, then monitor it and expect it to crap out for some reason occasionally. At least expect your backup computer to run out of space eventually.
Using WOL you need to account for booting time, including the occasionally slower than average boot time.
How will you handle software (OS) updates? You don’t want your backups to live in an unsafe environment, do you?
Backup computers in home labs typically run on older hardware - I once had set one up to turn on every 4 hours until the interface to set the wakeup time in the BIOS was deprecated in Linux (yes, the computer was THAT old). Certainly an unusual way to break backups.
My recommendation is to use open source software and keep things as simple as possible.
I recommend finding a Linux distro for the backup computer you’re comfortable with. Kudos if you can drop the window manager. There is in principle nothing wrong with staying on Windows or using another OS, e.g. FreeBSD in TrueNAS.
Think about what exactly you want to back up. Personal files? Copies of your media collection? Your computer setup? “Everything”?
As important as figuring out what to backup, is thinking about (and testing) the process of restoring a backup. Will this be an all-or-nothing or a file-by-file approach?
In principle, you probably don’t want to run a complete backup every time - because this can take a very long time.
After an initial complete backup, you will want to implement incremental backups.
There is software that does that, but often these incremental backups are stored in a proprietary file format or at least in a format that requires the backup software to work for a restoration.
Alternatively, you can implement a file based backup. A popular software is rsync (Linux has a tool called rsnapshot to simplify backups). This will implement incremental backups stored efficiently on the backup computer and is presented in a filesystem for easy recovery with OS-level tools.
Finally, you can use file system tools, e.g. sending (incremental) snapshots or COW filesystems, such as zfs or btrfs. The positive is very fast incremental backups, more (btrfs) or less (openzfs) OS level functionality (no special backup software required). Can be intimidating at first.