Btrfs Assistant 1.0 is coming, testers needed

dalto · September 12, 2023, 3:31pm

Yes. That would basically destroy everything and start over.

milkytwix · September 12, 2023, 4:54pm

The “change the mountpoint” version does not work:

sudo mount /Virtual
mount: /Virtual: mount point does not exist.
       dmesg(1) may have more information after failed mount system call.

fstab:

cat /etc/fstab |grep irtual
UUID=178ba18b-3faa-4b38-b7e5-5b0f5d2f285c /Virtual       btrfs   subvol=/@virtual,defaults,noatime,compress=zstd 0 0

And although /virtual is unmounted I cannot rename it…

dalto · September 12, 2023, 4:55pm

The mountpoint needs to exist. I thought you were already using it so I didn’t add that point.

You can just create it:

sudo mkdir /Virtual

milkytwix · September 12, 2023, 5:06pm

A, okay, I thought I can rename the /virtual… Creating /Virtual and deleting /virtual did the trick. Man you are the best…

milkytwix · September 13, 2023, 10:24am

One last question before I create the mirror: My current drive 1 has two partitions on it, 1000MB FAT32 for Grub, the Rest as BtrFS for EndevourOS. When I now remove the ext4 partitions on drive 2 will I create one BtrFS partition for the whole drive two? Or do I duplicate the whole partition scheme? I could then backup the grub partition regularly, and if drive1 fails, write this backup to grub partition on drive 2 and boot the machine up - at least in theory…

milkytwix · September 14, 2023, 1:11pm

So, I just removed everything from device two and then just added it via:

sudo btrfs device add /dev/nvme1n1 /
sudo btrfs balance start -dconvert=raid1 -mconvert=raid1 /

And it worked… somehow?

lsblk -f
NAME        FSTYPE FSVER LABEL       UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
zram0                                                                                    [SWAP]
nvme0n1                                                                                  
├─nvme0n1p1 vfat   FAT32             BDEB-C54C                             997,5M     0% /boot/efi
└─nvme0n1p2 btrfs        endeavouros 178ba18b-3faa-4b38-b7e5-5b0f5d2f285c    1,5T    17% /var/log
                                                                                         /var/cache
                                                                                         /home
                                                                                         /Virtual
                                                                                         /
nvme1n1     btrfs        endeavouros 178ba18b-3faa-4b38-b7e5-5b0f5d2f285c

Am I good?

dalto · September 14, 2023, 1:16pm

Yeah, it is really easy.

You can verify and see more detail with:

sudo btrfs filesystem usage /

milkytwix · September 19, 2023, 3:04pm

Well, I have this issue, which started right when I created the Raid1 (did the balance). Could be a seperate issue, most probably is, but nevertheless, it started when I had the raid up and running. So my question is: How to “unraid”?

BluishHumility · September 19, 2023, 4:04pm

You can remove a device just as easily as adding one, see here:

https://wiki.tnonline.net/w/Btrfs/Adding_and_removing_devices

Use btrfs device remove <device-or-devid> <mountpoint> to remove a device.

Be aware of the following limitations:

Limitations when removing devices

All of the data must fit on the remaining devices.

It is not possible to go below the minimum number of devices given the data or metadata profile used. For example on a 3 device RAID1 filesystem it is possible to remove 1 device but not 2.

See more about the limitations of different profiles on the Btrfs/Profiles page.

This is from the linked profiles page:

On a single disk filesystem the default is SINGLE profile for data and DUP for metadata chunks.

Prior to removing the device, you will have to convert the filesystem back to these defaults (because a single device cannot be RAID1).

sudo btrfs balance start -dconvert=single -mconvert=dup /

Then remove the device.

sudo btrfs device remove /dev/nvme1n1 /

milkytwix · September 19, 2023, 5:21pm

Worked like a charm. Aaaaand as expected, did not solve my problem…

tuxaholic · October 15, 2023, 9:50am

What’s the difference between the “Enable timeline snapshots” checkbox in the “Snapshot retention” panel, versus the “Snapper timeline enabled” checkbox in the “systemd unit enabled” area?

Also, if one disables “Snapper cleanup enabled” in the “systemd unit enabled” area, does that mean that the numbers for hourly/daily/weekly/etc will stop being “trimmed” and allowed to grow?

dalto · October 15, 2023, 2:05pm

The former enables timeline snapshots in Snapper. The latter enables the systemd service. Both need to be enabled for them to work.

Yes. Unless you run snapper’s cleanup routine some other way, the snapshots will never have any of their retention enforced and, as a result, never be removed.

If you enable timeline snapshots, having this disabled would be…bad.

tuxaholic · October 15, 2023, 4:43pm

Thanks, I’ve been reading the Arch wiki on this and have created a mapping. I will post it here in case it’s useful for anyone in the future:

“Snapshot Retention” section

Run command sudo snapper get-config to see these settings after saving, they map to:

“Enable timeline snapshots” check box = TIMELINE_CREATE
“Hourly” setting = TIMELINE_LIMIT_HOURLY
“Daily” setting = TIMELINE_LIMIT_DAILY
“Weekly” setting = TIMELINE_LIMIT_WEEKLY
“Monthly” setting = TIMELINE_LIMIT_MONTHLY
“Yearly” setting = TIMELINE_LIMIT_YEARLY
“Number” setting = NUMBER_LIMIT

The “Number” limit applies to non-timeline snapshots, typically boot/manual/pre-/post- type snapshots.

systemd Unit Settings section

The meaning of each check box is:

Snapper timeline enabled: Turn timeline snapshots on/off (check with systemctl status snapper-timeline.timer)
Snapper cleanup enabled: Turn periodic cleanup service which deletes older snapshots on/off (check with systemctl status snapper-cleanup.timer)
Snapper boot enabled: Turns taking snapshot at boot on/off (check with systemctl status snapper-boot.timer).

Cheers!

HMS_Endeavour · October 19, 2023, 6:25pm

Hi @dalto Its that time again to rebuild. Can you recommend any tutorial to setup grub-btrfs with BtrFS Assistant on a default BtrFS setup? I don’t really want to use Timeshift.
Thanks.

Brtza · October 19, 2023, 6:27pm

Just wanted to ask him something similar . I did manual partitioning , with separate btrfs (/) and btrfs (Home). but after install it seems that is everything in root (/) …heh

BluishHumility · October 19, 2023, 6:52pm

He already laid out the steps for you in this same topic: Btrfs Assistant 1.0 is coming, testers needed - #12 by dalto

dalto:

After installing with btrfs selected as the filesystem:

Install snap-pac, snapper, grub-btrfs and btrfs-assistant

Open Btrfs Assistant

Switch to the “Snapper Settings” tab

Click on the “New Config” button

Name the config “root” and choose / as the mountpoint

Click the “Save Config” button

Set the retention limits you want. In particular you probably want to lower the amount of “Number” snapshots that are retained. These are the snapshots that are taken when pacman runs. It takes 2 snapshots in each run so something like 10 is probably a more reasonable number. That will give you 5 pacman runs. The timeline settings are entirely left to your preferences.

Select the checkboxes for snapper-timeline and snapper-cleanup and click the “Apply” button to the right of the checkboxes.

Add support for booting off read-only snapshots

Edit /etc/mkinitcpio.conf and add grub-btrfs-overlayfs to the end of the HOOKS section

Rebuild your initrams with sudo mkinitcpio -P

Make sure the snapshots are added to the grub menu. There are two easy ways to do this. Only pick one of them

Enable the grub-btrfs.path systemd unit

Install snap-pac-grub from AUR

I think that is pretty much it. Of course you can also create snapper configs of other subvolumes if you like. For example /home. You just need to create a config for it in the “Snapper Settings” tab. Make sure you enable timeline settings for it as only snapshots of “root” are taken during pacman operations.

No need to do the mkinitcpio step of course, if you have a recent ISO that is using dracut. Just rebuild the initramfs the dracut way:

sudo dracut-rebuild

Although that reply was from April 1st (2022), I do not think it was an April Fools’ joke.

Brtza · October 19, 2023, 6:54pm

Thanks . But what about partitioning. What is the proper way to do that ?

BluishHumility · October 19, 2023, 6:55pm

No special partitioning is needed. Just make sure you use Btrfs for the filesystem. The default partitioning offered in the Calamares installer is fine.

dalto · October 19, 2023, 7:24pm

If you want to be able to boot and restore root snapshots there are a few considerations:

/ should be mounted on a subvolume. It doesn’t matter what it is named.
The / subvolume should contain everything you need to restore. i.e. don’t put /usr on it’s own subvolume or when you restore / it will be a problem.

Other than that, the default Calamares layout will work just fine.

HMS_Endeavour · October 19, 2023, 7:25pm

ok, I have some memory loss at my age. Anyway, I don’t have a mkinitcpio pckage installed.