With the amount of data I have, 16 to 20 hours seems normal.
So I have an extra drive same capacity as my main qubes drive. Can I setup RAID or some type of disk mirroring so backup can happens passively in the background over time?
I don’t know how to do this. I’m guessing some config through dom 0? The ideal for me would be if the main drive fails, I just boot to the mirror drive and swap out the failed drive and that fills again. 16 to 20 hour backups are not possible for me, so RAID 1 or some linux variant would be better. Can you advice?
Okay you’ve likely properly scared me away from raid. Getting copies of corrupt files from failed harware was a worry. As a solution to slow backups, do you know if its possible to “suspend” the computer during a backup and have the backup resume until completion without issue? This would solve my problem. Due to my lifestyle it’s just impossible to leave my computer open for more than 10 hours.
Think about this another way.
How can you reduce the time taken by those backups?
What qubes are you backing up? Do you need them all on every run?
Do you need to backup all the data in the qubes on every run?
I have no idea how your data is distributed between your qubes - you
haven’t said.
Let’s say you have 1TB data in a “data” qube, and small amounts in your
other qubes. Why would you want to backup all that data on every run?
The obvious solution is to:
Run a full backup of the data qube.
Run daily incremental backups (using a tool like rsync) of data that has
changed to some other qube, “increment”.
Back up the increment qube daily.
Alternatively, make use of that extra disk, and create a new “mirror”
qube.
Daily sync your data from “data” to the “mirror” qube.
Run a Qubes backup of the “mirror” qube, while working on “data”.
Just because you have a Qubes tool that backs up all your data, that
doesn’t mean that it is the best tool to use.
The advantage is that you will have a secure, encrypted backup which
you can restore in one fell swoop.
The first method requires at least 2 restore operations, but at
considerable saving in time and disk space. It will also allow you to
cherry pick changes from specific dates, so that if a file has been
accidentally deleted or modified, you would be able to go back to that
point in time. That, of course, is what a backup is for.
You’re correct. Going through these qubes, there’s tons of test vms I created which I no longer use. I’m starting to see at least 500gb I don’t need.
Considering rsync… I’ve just started reading about it, if I want to back up let’s say 10 different qubes, and automate this, is the only way through dom 0 and chron jobs? Chron jobs I’m also new to but will learn if it’s the right way. However I’d like to never touch dom 0 if possible. Thanks!