These are public posts tagged with #btrfs. You can interact with them if you have an account anywhere in the fediverse.
So, I’ve basically reinstalled the system from a USB drive and am now restoring #backups. Luckily, I backed up all the #Dotfiles and configurations and so the #desktop is behaving as before, which is a big relief. On the up side, it’s booting a lot faster and the #disk has #btrfs #partition, rather than #Ext4.
I have a machine running #btrfs RAID1 with two SATA SSDs.
If I were to replace one of them with an NVMe SSD, would that improve performance, or would it be bottlenecked by the remaining SATA SSD?
(I know mdraid has a write-mostly mode, but I don't want to use it because (1) no bitrot correction and (2) I'd rather live with a performance hit than take my chances with a fragile tower of storage layers.)
Je dois changer de disque (je remplace mon SATA 350 Go par un SSD Crucial 1 To : https://www.debian-fr.org/t/migration-sur-ssd/91101 ) et je voulais partir sur du LVM/BTRFS.
On me dit que LVM/BTRFS c'est me compliquer la vie pour rien, sur une machine d'usage ordinaire ext4 suffit largement et en tous cas BTRFS fait double emploi avec LVM.
Un avis ceux qui utilisent #BTRFS ?
Bonjour Debian. J’ai un truc à installer sur mon PC…
debian-fr.orgMy experience with #FlashDrives recently has been mixed. I have no problem in encrypting them with #LUKS, using #cryptsetup or with formatting a partition with #Btrfs, for instance, using #gparted and doing other tinkering with #Gnome #disks. But the problem has been with the actual drives themselves. The cheaper ones seem to have quite a few bad sectors, etc. and so they’re not really reliable for medium term storage.
1/2
Power on hours: 143724
Reallocated sector count: zero
#Btrfs total checksum mismatches: zero
Btrfs total read/write errors: f**king zero
Slow though they may be, hard disk drives are amazing. #SSD s are cool and all, but how often do you find one that's old enough to drive a car and still functioning perfectly?
Using the following command I verified that I already have all the data elsewhere:
#btrfs inspect-internal dump-tree /dev/sdc1 | grep name:
So I'm nuking the partitions and starting from scratch...
Just hit https://github.com/raspberrypi/bookworm-feedback/issues/107 with #btrfs
> BTRFS error (device sdb1): sectorsize 16384 not yet supported for page size 4096
I can't mount the filesystem since it was created on a Pi kernel with 16k page size, and now I'm on 4k kernel due to compatibility. I can no longer access my data :(
I'm sad...
This issue will be a public tracker for 16K pagesize…
GitHubAnfang des Jahres hat @ammoniumperchlorate einen Vortrag darüber gehalten, wie wir unsere Mastodon-Instanz rheinneckar.social dank #btrfs und #NixOS mit nur wenigen Sekunden Downtime auf einen neuen Server umgezogen haben. Definitiv sehenswert!
Enjoy the videos and music you love, upload original…
www.youtube.comCette semaine #timeshift et les snapshot #btrfs m'ont permis d'éviter une réinstallation complète du système et d'y passer mon week-end.
Que s'est il passé ?
J'ai tenté une mise à jour partielle de ma #Manjaro avec yay en n'installant pas une librairie qui n'était pas compatible avec une application aur.
Résultat j'ai tout cassé, plus aucune appli ne fonctionnait. Impossible de démarrer l'OS.
Dans GRUB, j'ai listé les snapshot et modifié le boot pour démarré sur le dernier fonctionnel.
Thanks @vkc
<Start/tip nobody asked for>
For those who want something close to Debian testing, #siduction (by default #KDE #Plasma6) might be worth a try. It is unstable (Codename: Sid), thus before testing. This means it is tested but, it is certainly not as stable as testing.
But here is the twist. Use it with #btrfs or #timeshift and #ext4 to have efficient tools for a rollback once it breaks (and it will break sporadically) and you should be good.
<End/tip nobody asked for>
Eigentlich wollte ich den #Linux Kernel 6.14 testen, aber #initramfs, #btrfs und #nvidia sagen nein.
I'm debating writing a little shell wrapper to list and delete #btrfs snapshots, since they're generally just subvols with predictable names and this kind of thing should be more immediately intuitive than that from CLI.