this post was submitted on 07 Sep 2023
1318 points (96.0% liked)

Memes

45731 readers
1271 users here now

Rules:

  1. Be civil and nice.
  2. Try not to excessively repost, as a rule of thumb, wait at least 2 months to do it if you have to.

founded 5 years ago
MODERATORS
 

OC for you.

you are viewing a single comment's thread
view the rest of the comments
[–] Hovenko@iusearchlinux.fyi 3 points 1 year ago (1 children)

YOu can set those pre/post snapshots automatically and not really pay attention. I think OpenSuse does that by default if you install your root on btrfs. They even have an OS version called MicroOS which does a cool thing with snaphosts. Basically if your system does not boot after update it will revert automatically to previous snapshot, or you can pick a snapshot to boot into manually from grub menu. Bit it is quite a different thing than your usual linux distro as it uses read only root FS.

[–] 01189998819991197253@infosec.pub 1 points 1 year ago (1 children)

Ooh! Ok cool! I've never used btrfs. How does it compare to ext4?

[–] Hovenko@iusearchlinux.fyi 2 points 1 year ago (1 children)

Hmm... it's complicated since it's a very opinionated topic. There will be people who like it and there are people who will kill your entire family because you mentioned it :D The benefits depend on what are you using it for. The difference is mainly in features it provides. I am using it mainly for snapshots, subvolumes and raid support. Raid is stable only for jbod, 0, 1 and 10. 5 and 6 are still having issues. Raid is self healing if I am correct. So if there are some checksum errors it will repair stuff from working mirror.

You need to do some research before using it. For example since it is cow FS it is not good for databases or VM images. But you can turn off the QCOW feature for specific directory. The other thing is space usage. Due tyo snapshots and cow feature 'df -h ' will not show exactly correct usage data. You have 'btrfs filesystem df -h ' instead. If you kill up the disk it can be more bothersome to clean it up afterwards.

I would say, try it out and decide yourself. There are definitely videos online that explain it better than I did just now.

[–] 01189998819991197253@infosec.pub 1 points 1 year ago (1 children)

Sounds like btrfs is still more on the bleeding edge side of things. Sounds pretty cool for a tester laptop. I installed OpenSuse with it last night, but forgot to format the drive first, so it installed on a small partition. I would just rearrange partitions, but the manpages state that it can cause issues with btrfs, so a reinstall it is haha

[–] Hovenko@iusearchlinux.fyi 1 points 1 year ago (1 children)

Oh cool. Yeah. The update pre/post snapshots are enabled by default so you need a bigger partition. I think they recommend 60g? But in the worst case you can turn them off.

Regatding bleeding edge thing. I think it is pretty stable for general use. SLES (opensuse enterprise) is using it as default for root fs enterprise deployments. I don’t think they would risk finantial loses using unstable filesystem.

I can imagine the problems will probably appear on large scale deploymets, maybe with deduplication. For example I am using btrfs on my home server in raid10 array… for abou 6-8years now. I had issue only once so far which was due to me being a dumbass.

Oh, that's really good to hear! I plan on reinstalling OpenSuse today, after I delete all partitions from the drive. I daily Fedora, and like gnome fine, but kde is what started me on Linux, and using OpenSuse yesterday was so nostalgic. I forgot how much I enjoyed kde.