Hi to you all,
Are there any new “views” on btrfs/xfs and ext4?
I always worked with ext4 and I’m curious what btrfs would bring me.
More stabillity? More speed? More …?
Grtz
Kage
More bells and whistles . There are option that are LVM like and RAID like then there is snapper that lets you roll back changes. Seems stable at least in simple configuration. Speed is more or less the same depending on what tests you run.
I guess that depends on what counts as new.
More stabillity? More speed? More …?
More headaches – does that count?
I’ve tried “btrfs”. I originally tried it on a beta release of 13.1. And I went back to “ext4”.
Most recently, I used it on Tumbleweed for a while. But I eventually reinstalled using “ext4”.
I think it varies from user to user. It depends on what you are looking for.
The main benefit of “btrfs”, as far as I can see, is the ability to roll back to an earlier snapshot. I don’t see myself ever using that. If I badly mess up my system, I would prefer to do a clean install (keeping “/home” which is a separate file system).
The main negative, for me, is that I sometimes like to mount one system from another. So, while running Leap 42.1, I like to be able to mount Tumbleweed (installed separately but on the same box), so that I can tweak something. But “btrfs” makes that harder, because I have to mount all of the subvolumes to make that work.
In theory, it ought to be possible for system updates to create a new snapshot, and update to that new snapshot. In the meantime, I continue running the non-updated system in the current snapshot, with no conflicts between old versions and new updated versions. When I boot to the new snapshot, I see the updated system. But somebody would have to do a lot of work to implement this properly.
Hi
No issues here with either btrfs or xfs as a filesystem for a number of years now.
I’ve used rollback a couple of times but that is snapper related, not the file system… 
which IMHO needs some config tweaks.
Since running Tumbleweed and if you get numerous updates you do need to look at running the snapper cleanup cronjob manually as well as the btrfs-balance one. I find that with a 40GB / it’s using about 15GB and allocates about 24GB so at a pinch you could get away with 30GB.
You can always disable snapper/snapshots as well…
btrfs seems to write more data to a disk than ext4.
So if you have a SSD and care about the wear of that then …
Hi
I never worry about it… my latest SSD SanDisk Extreme Pro 240GB has a 10 (limited) year 50GB a day warranty, Have OCZ’s with thousands of hours running btrfs all running fine. my oldest one a OCZ Agility3 60GB is at almost 32K hours and still shows no wear…
I also have a OCZ Vertex4, OCZ Vector and a Crucial SSD, no wear in daily use…
Well it writes little more data but it does keep more data. Blocks that are rewritten are not immediately reused but preserved and added to snapper’s lists when a snapshot happens. But that is snapper. Turn it off and BTRFS will not use any more data then any other FS, In any case modern SSD have sufficient life and redundancy to weather the small amount of extra data writes associated with snapper block lists (snapshots)
IMO the main supposed advantage/feature of BTRFS is that it’s supposed to be “self-healing” ie you should’t have file system damage which would require manual fsck as often. BTRFS is supposed to fix problems invisibly to the User without prompting.
Then again,
I haven’t had to fsck an ext4 in about… 8 years?
I can certainly remember earlier systems (running ext2? Can’t remember) and a particular system one guy dumped in my lap which would need to be manually repaired like clockwork every couple days or so.
Probably the most well known BTRFS is snapshots, although it’s also possible to run a snapshot utility on ext, it’s just not built into the file system. Although most people probably think of snapshotting as more a bother taking up space than useful, this is the kind of thing that the first thing you use it to save your bacon will change your mind completely.
There has been some critcism of ext in recent years that it’s getting too big/bloated, but I don’t share that view for now as long as it’s reliable and performant.
TSU
OK guys
Many thanks for all the input.
I guess I’ll stick to ext4.
Grtz
Kage