I wonder if #btrfs will ever have per-subvolume profiles.
This would, among other things, allow one to have a swap file on a RAID array. Just add a subvolume with a non-RAID profile, and you can put a swap file on it.
Administered by:
#btrfs
7 posts6 participants0 posts today
#linux
#filesystem
#justLearned
#BTRFS
#student
I saw BTRFS when I called gmesg. I turned off Bluetooth on machine before hand, and I was alarmed... Blue Tooth Radio Frequency... ? FORTUNATELY, after I searched info, I found and learned this is a type of file system.
From https://btrfs.readthedocs.io/en/latest/ :
"BTRFS is a modern copy on write (COW) filesystem for Linux aimed at implementing advanced features while also focusing on fault tolerance, repair and easy administration."
btrfs.readthedocs.ioWelcome to BTRFS documentation! — BTRFS documentation
> "Butter FS" est l'un des systèmes de fichiers les plus avancés disponibles aujourd'hui sous Linux. Il est moderne, repose sur des principes qui le rendent extrêmement fiable et propose de nombreuses fonctionnalités très intéressantes.
> Malgré cela, il se traîne une mauvaise réputation qui freine son adoption. Mais cette réputation est-elle vraiment justifiée ?
**Btrfs : révolution ou catastrophe ? Où en est-on aujourd'hui ?** par @FLOZz (2022)
https://blog.flozz.fr/2022/05/22/btrfs-revolution-ou-catastrophe-ou-en-est-on-aujourdhui/
FLOZz' Blog · Btrfs : révolution ou catastrophe ? Où en est-on aujourd'hui ?Btrfs (prononcez « Butter FS ») est l'un des systèmes de fichiers les plus avancés disponibles aujourd'hui sous Linux. Il est moderne, repose sur des principes qui le rendent extrêmement fiable et propose de nombreuses fonctionnalités très intéressantes. Il est même utilisé par défaut par plusieurs distributions Linux. Malgré cela, il se …
Replied in thread
mit 
@sun @Lars I decided against using btrfs for another couple years this past week after reading plenty of recent reports of its failure to recover from errors, along with its longstanding reputation of being unable to recover from many types of errors, see also https://www.usenix.org/system/files/atc19-jaffer.pdf
#btrfs would be a lot better if it *handled* errors rather than detecting them (which it's great at) and then blowing the fuck up.
btrfs mention
Replied in thread
@bluelupo
Ich nutzte sie gerne. Aber ja, es ist eine distribution bei der #timeshift oder #btrfs sehr empfehlenswert sind.
Next change to my #selfhosting setup is moving an install on one SSD (root) + one HDD (storage of media files) to a single SSD with #btrfs partitions on top of #LVM
Should make the server whisper quiet
Continued thread
anybody know if #btrfs has improved this situation with regard to filesystem recoverability after medium errors common on consumer SSDs?
Continued thread
I made this crontab to keep your #btrfs based #gitlab runner happy without to much maintenance.
Should keep you from running out of disk space.
5 0 1,15 * * smartctl -t short /dev/sda
5 1 1,15 * * btrfs scrub start /
5 4 1,15 * * btrfs filesystem defrag -r /
5 0 2,16 * * btrfs balance start -musage=50 /
5 4 2,16 * * btrfs balance start -dusage=10 /
5 0 3,17 * * docker image prune -a
Ran out of disc space on #btrfs now praying to the rebalance gods after adding a USB drive to the pool.
Not how I wanted to spent my morning..
Continued thread
Can I just add any old hard drives (any size) into a #btrfs raid-1 pool? If I can just snag hard drives from wherever, and grow the pool as I scrounge more drives, that would be ideal
Behold the latest episode of "Tuning the Titanic's Deck Chairs," featuring Anand Jain's epic saga of #Btrfs allocator hints! Because clearly, chunk allocation was the #performance #bottleneck we all needed solved. 
Meanwhile, #sysadmins everywhere eagerly await the next riveting installment of "PATCH 0410" to complete their collection. 
https://lwn.net/ml/all/cover.1747070147.git.anand.jain@oracle.com/ #TuningTheTitanicsDeckChairs #PATCH0410 #TechSaga #HackerNews #ngated
lwn.net[PATCH RFC 00/10] btrfs: new performance-based chunk allocation using device roles [LWN.net]
Continued thread