anyone tested Mint KDE btrFS snapshots ?

[chan] linux
Dec 2 22:57

mint now ships with a btrfs snapshot GUI so in theory u can rollback after a package install fucked up ure system usable in practice ?

[chan] linux
Dec 2 23:14

Why use mint KDE when there is Kubuntu? The only reason I see for using mint is the Cinnamon DE. I dearly wish there were a genuine Ubuntu flavour with Cinnamon desktop, then I wouldn't have to resort to Mint. (But the only way is through installing extra packages which don't really work). Scrap Ubuntu MATE or XUbuntu for the sake of "CinUbuntu", and I wouldn't miss Mint one bit! Mint has no advantage over Ubuntu, except the Cinnamon DE. There are only disadvantages, like no easy way to d-u, or Kernel updates aren't installed automatically once you've chosen a different kernel and don't stay with the most current version of mint. That's all bullshit, cobbeling things together, but not a concise way of doing software updates. Again, if there were an official flavour of Ubuntu with the Cinnamon desktop, it would be *the* best OS there is!

[chan] linux
Dec 3 02:37

Yes, its a nice thing I have a package apt-btrfs-snapshot installed (ubuntu) which automatically creates snapshots every time I do something with apt, like install, update, remove packages. So in case of fuckup I can load ubuntu from usb stick (or from the console instead of gui if its still available), and do the sudo mount /dev/sda1 /mnt/ cd /mnt/ ls and I will see list of directories, where each directory is actually a snapshot - @ @backup-2017-12-01 @backup-2017-12-02-13:37:00 @ is root directory, it contains current files, I can rename @ to something else like @fucked, and rename "@backup-2017-12-02-13:37:00" to "@" so this particular snapshot will be loaded next time ubuntu starts. This is easiest backup restore ever. If old snapshot works fine then delete @fucked later. Also I can open each snapshot directory to view/copy/change older files, or do "sudo btrfs subvolume delete -C @backup-2017-12-01" to delete old snapshots (they do take some disk space, and must be deleted manually). Also you can just change /etc/fstab parameters to select particular snapshot to load system from. These snapshot folders will show you full list of files in file managers, but deep in the filesystem they actually contain only the changes comparing to the parent snapshot. But you can delete parent snapshot, and all child snapshots and files will still be fully available, which is some sort of magic. Disk check and defrag tools for btrfs are very beta, be cautious with subvolume deletion. Restart your pc asap if you delete root snapshot, as these changes must be commited asap, or your filesystem will become corrupted in case of unexpected shutdown like power failure. I had this issue once, couldnt fix it. btrfs natively supports all kind of raid arrays, just specify all devices names and raid type when creating btrfs filesystem. This might be great for raid0 or raid1. Don't use btrfs raid5 or raid6 tho, they are very beta, and have bugs which can destroy whole filesystem in certain conditions.

[chan] linux
Dec 3 10:42

In theory btrfs is pretty cool; in practice your btrfs can be a lot of trouble: - you run into a disk space allocation deadlock ("no space left on device" despite using only a fraction of disk) - your vfs accesses are exponentially slower than with ext3 or ext4 (compare the time it takes to unpack and repack a large tarball or compressed git repo, e.g. the linux kernel; or just run a db) - you lose data: files disappearing or being unreadable (despite the disk/block device being 100% ok) The speed/performace trouble even happens on a really fast machine (both ssd and hdd); the data loss/allocation trouble happened every time i tried to give btrfs another chance. Any of these troubles never happened to me with ext3/ext4. The only data I remember losing with ext4 was either by REALLY defective hardware or locking myself out of the full disk encryption system. And be careful with cheap-ass USB sticks: some of them are already defective out of the box (as in lots of bad blocks) and no file system can cope with that (if the hardware doesn't even report a "write error", but says "write successful" when the data wasn't written correctly). A year ago I had bought a batch of those (EUR 16 per 64 GB each) all of them were defective. Good thing I had kept the bill and could claim warranty.

[chan] linux
Dec 3 17:09

> you run into a disk space allocation deadlock ("no space left on device" despite using only a fraction of disk) You can have multiple root and child subvolumes, and each will take some space, but file manager will show disk space taken by currently mounted subvolume only sudo mount /dev/sda1 /mnt/ and check disk usage in /mnt/ , it will be much larger comparing to / in case of multiple subvolumes. You can run out of space on the disk even if currently loaded subvolume is not even full yet. Run something like sudo btrfs subvolume list -pat /mnt/ to show all subvolumes on your disk

[chan] linux
Dec 4 16:23

Because ubuntu is a bad distro. Better use debian.

[chan] linux
Dec 5 02:34

That happened while using only one subvolume.

[chan] linux

Subject Last Count
hi there Dec 13 17:13 2
linux distros Dec 10 18:05 1
The Stallman Tax Dec 9 14:48 1
hello Dec 9 08:39 2
OMEGA release 39 is available for download Dec 5 23:53 2
the mammoth is served on the www web : Dec 5 12:51 1
anyone tested Mint KDE btrFS snapshots ? Dec 5 02:34 7
cool lap Dec 4 16:23 1
fuck me, I went back to MX distro for giggles Dec 2 22:52 11
the fuckers at debian Dec 2 14:37 2
BM w/ GUI : Jupyter BM notebook Nov 30 12:13 4
easy mailer , just add MX line in DNS Nov 30 10:55 1
onion mail Nov 29 00:56 1
How to build the pow for linux? Nov 27 21:09 4
encryptpad Nov 26 03:31 1
Alex Jones Wanker Nov 20 21:53 1