Hello people,
With Windows, I used Macrium to back up disks and restore them with good options for resizing partitions.
I use Vorta, and I use BorgBase for my home partition online.
I want to do a full disk backup with my laptop.
I have an encrypted partition with a Btrfs partition.
In the past, I used Clonezilla, and I’ve seen it support LUKS and Btrfs.
So, I attempted to do a backup booting from a USB pen drive.
I’ve seen that Clonezilla can back up my disk in two ways when dealing with the encrypted partition: in raw mode or opening the LUKS volume.
The first mode is, of course, very long.
So I wanted to attempt the copy, letting Clonezilla open the LUKS volume and copy the content.
I’d like to know if someone else has copied and restored the disk this way.
I searched for an alternative to Clonezilla and always got “dd” as a response.
Please let me know if someone knows an alternative backup system with LUKS and Btrfs support and incremental backup support.
Mmh, that’s interesting.
The possibility of backup snapshots on my USB/NAS is good.
I’m looking for an entire disk backup solution because if I “backup” snapshots, I have to recreate the disk structure before restoring them or something, I think.
That is pretty much what Clonezilla do.
With Fileazilla, I can use raw mode or not.
The raw mode is not suggested because I know it’ll take hours for a full backup.
So, Filezilla will open the LUKS volume and copy all data in btrfs.
Said that,
I don’t know how long it will take to use Rsync or similar for a full restore.
I don’t think that rsync will take care of recreating the partition structure on my disk.
However, I’ll look at what you suggested, of course.
I don’t see the point of copying the partition structure. Just copy the directory hierarchy and in case of catastrophic failure, recreate the same partition structure and restore directories using rsync.
I’m a newbie, so I’m a bit in trouble. And I don’t know how to proceed in a lot of situations.
My current disk structure is as follows:
zram0 252:0 0 8G 0 disk [SWAP]
nvme0n1 259:0 0 953,9G 0 disk
├─nvme0n1p1 259:1 0 600M 0 part /boot/efi
├─nvme0n1p2 259:2 0 1G 0 part /boot
└─nvme0n1p3 259:3 0 952,3G 0 part
└─luks-db705de3-ceda-484c-b529-0f7ef7a7c206 253:0 0 952,3G 0 crypt /home
So, I have an external backup, and my disk is broken.
I change the disk. Then what?
Do I’ll boot with Gparted, and then recreate the same structure, and then restore with rsync?
Do you intend to restore every single thing including your Fedora install or only your files? I assume your files are all in your home.
If you are ok with reinstalling Fedora after a disk failure, you can recreate this same layout in the installer, connect the backup disk then rsync while conserving file attributes.
I prefer rsync over btrfs for backups since it allows for backup disks other than btrfs.
It supports full and incremental backups.
On my main desktop machine it runs every hour so that i can get back any file I accidentally change or delete.
I have a file server, also fedora, that its stores the backup data on.
This is all automatically run as systemd services.
Every now and again I copy all the backup data from the file server on to an encrypted external disk that I store away from the house.
On older systems, I used to do full backups with rsync as per the Arch Linux Wiki.
On systems where I have btrfs or zfs as root, I use the send|receive functions of those filesystems and that works very well.
The downside of send|receive is that you need the same filesystem on your backup destination (although btrfs can “send to a file” but IMHO you don’t want this kind of complication in a backup workflow). Also doesn’t work well with object storage if you want to add that to have a 3-2-1 scheme.
Now, this is probably not the answer you’re looking for but I have moved away from full disk backups as this is not considered best practices (anymore). I have not looked back.
The main reasons are that the full disk backups are very heavy, and restore is both destructive and more complicated.
I have a comprehensive Ansible playbook where I write down the system’s config and changes made to it.
To restore, I just need to reinstall Fedora without touching the homedir and deploy the Ansible script.
Or wipe everything, install a fresh Fedora, restore the homedir and deploy the Ansible script.
Restoring from scratch to a fully up and running system takes <2 hours including LAN transfer times.
This is probably way more effort than you want to invest but it’s possible to do block device level back up using ceph rbd or rgw. That assumes you already have, or have the resources to set up, ceph, which would be way overkill to support one box.