Checking my disk configuration

Hi all.

A few days ago I installed Fedora 42 on my laptop and the disks looked like this:

NAME   FSTYPE FSVER LABEL  UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
sda                                                                            
├─sda1 vfat   FAT32        2FD9-1F5E                             579,5M     3% /boot/efi
├─sda2 ext4   1.0          35a8c2ca-fb27-4611-9590-d19a08a8a9ed  535,6M    38% /boot
└─sda3 btrfs        fedora d911e50b-02f2-4615-9443-17e8a63ac351    1,1T     4% /home
                                                                               /
sdb                                                                            
└─sdb1 btrfs        fedora d911e50b-02f2-4615-9443-17e8a63ac351                
zram0  swap   1     zram0  ccfd2fae-8ada-48ed-8d6a-5395cdc6e689                [SWAP]

#
# /etc/fstab
# Created by anaconda on Fri Jun 27 09:28:03 2025
#
UUID=d911e50b-02f2-4615-9443-17e8a63ac351 /                       btrfs   subvol=root,compress=zstd:1 0 0
UUID=35a8c2ca-fb27-4611-9590-d19a08a8a9ed /boot                   ext4    defaults        1 2
UUID=2FD9-1F5E          /boot/efi               vfat    umask=0077,shortname=winnt 0 2
UUID=d911e50b-02f2-4615-9443-17e8a63ac351 /home                   btrfs   subvol=home,compress=zstd:1 0 0

I understand that /dev/sdb1 is not being used at all, although I tried to install the spin of sway on it and he told me it was part of another volume, so I didn’t dare.

Today I extended the memory to 16 GB and need guidance to:

  • Clearly view volumes and subvolumes btrfs
  • Expand the zram to 17GB to be able to hibernate
  • Create a btfrs subvolume for snapshots
  • Learn how to use timeshift to make and retrieve snapshots

Any reference or web that helps me study all this will be of great help.

If you look at the UUIDs for sda3 & sdb1 you might understand that both partitions are contained in the same btrfs file system and are used as such.

You would need to remove sdb1 from that file system before it could be used for anything else.

Not possible in that way.
ZRAM is contained in RAM and is volatile so when you power off it disappears. To hibernate you would need to create a physical swap partition then use that for hibernation.

I do not use btrfs and do not hibernate so I cannot provide the details.

:pensive_face: Lack of attention on my part, thank you very much.

A very obvious misconception that demonstrates my clumsiness.

Thank you very much for correcting me

It is a very common mistake, and there is a lot to take in when configuring a system disk. I doubt you are more “clumsy” than most of us (I recently left a USB Linux backup drive connected when running a Windows update, something I know to avoid, and Windows added a small partition by overwriting an XFS partition).

You asked a question that others will have, so this topic may help them. I consider it a useful contribution to the community.

2 Likes

Yep you need to use

btrfs filesystem usage /

And this will show that there’s actually two device partitions making up one file system. It’s a multiple device btrfs. Depending on the result is should be possible to use btrfs device remove to remove one of them.

It’s important to note that this command starts immediately and doesn’t return to a prompt until it completes. During this time it’s migrating block groups from the device being removed to the device remaining. This could take minutes, hours or days, depending on the speed of the underlying devices and how many block groups need to be moved.

It is possible that the process gets interrupted: crash, power fail, runs out of space on the remaining device. So it’s important to double check that it completed. There will be message in dmesg as well as you can check btrfs fi us again.

Ahh, actually as I think about it, there’s the possibility the remove will be rejected. If this is a default Fedora installation, and both drives were chosen at installation time, the metadata block group profile will be raid1. Data profile will be single. The raid1 profile in a two disk configuration will prevent device removal because the minimum number of devices for raid1 is two.

So first you would need to convert the profile from raid1 to single or dup, and dup is preferred.

btrfs balance start -mconvert=dup /

Now the previous command will work and permit device removal assuming there’s enough free space to complete the operation. It will also shrink the file system, accounting for the loss of storage space due to the device removal. The magic will also be removed as a last step so that the kernel and blkid no longer see that it’s a btrfs device.

I do advise a back up before starting just in case. The device removal code is well exercised, it’s a variation on the balance code, moving block groups around is COW and at least with well behaved device firmware it’s crash safe - but better safe than sorry with a proper backup.

1 Like

Thank you very much @chrismurphy

The reality is that I was about to install with ext4 to take away the front but your message has inspired me to work to understand Btrfs and learn to live with him. It will be an effort but I think it will be worth it.

The truth is that I have to fight a lot against my tendency to reinstall every time I don’t understand something :flushed_face:

1 Like

I lost count how many reinstalls I did of Fedora (ages ago) because I didn’t understand something. :joy: Now I use Btrfs snapshots and rollback.

But commitment to a good backup and restore procedure is the path to calm testing. Just wipe the system, reinstall, restore. It’s surprising how many people backup but never test the restore and then disaster happens, and it’s bad news. So it’s worth getting as good at restore as backup, and frequent reinstallations got me pretty fixated on well tested disaster recovery.

Great truth :grin: