Bcachefs: “the COW filesystem for Linux that won't eat your data”

Post by jl6 » Thu, Mar 23, 2017, 7:11 AM UTC

noja wrote: post #40

zanny wrote: post #21

I've been using btrfs since about 2011, and I've stopped using ext4 / xfs / zfs everwhere since about 2014. From 2012-2014 it was mostly breakage every other month. From 2014-2016, it was semi-annual issues. For the last ~18 months I have had ~30 machines running btrfs with no issues, some servers, some personal computers. The release notes are boring, the bugs are boring, and to me its definitely in a state I would…

Me too. I love it, but only with raid-1 or no raid. btrfs as a filesystem will be dead within the next 12 months unless they fix raid-5 and raid-6.

Post by barrkel » Thu, Mar 23, 2017, 8:05 AM UTC

kazinator wrote: post #63

barrkel wrote: post #52

Earlier quoted context omitted.

Well, there are other reasons; you want to write code that operates on the data, and neither the code nor the data fits on a single machine - you have to target an abstraction which spans machines. Block storage is too low-level an abstraction. That isn't to say that using high performance block storage isn't still a win even when the redundancy is multiplied at a higher level. The higher level redundancy is also abo…

Block storage can be network-abstracted. Even virtual memory for that matter. Now ancient concept: https://en.wikipedia.org/wiki/Distributed_shared_memory

Post by kasabali » Thu, Mar 23, 2017, 10:41 AM UTC

feld wrote: post #15

std_throwaway wrote: post #13

Earlier quoted context omitted.

At first glance, the status page of btrfs look horrible: https://btrfs.wiki.kernel.org/index.php/Status The problem areas are mostly RAID and exotic features. RAID can be handled by a different layer and most users don't really need the exotic features. Judging from the media silence in the last months I'd say either people stopped using btrfs or it just about works good enough for everbody.

when the RAID is handled by a different layer you lose some very important integrity features

Post by nialv7 » Thu, Mar 23, 2017, 12:39 PM UTC

Zancarius wrote: post #77

tylerjd wrote: post #60

Earlier quoted context omitted.

ZFS is not GPL? I dunno, I like feeling safe when I do kernel upgrades knowing even if for whatever reason my ZFS module doesn't compile/work under the new kernel, I won't be left without a root FS. I have been running BTRFS for 5 years with very little issues, and enjoy not having to compile a new dkms module with every kernel.

ZFS with DKMS is a disaster, at least in my experience. Honestly, I can't recommend ZoL unless you're running a distro with relatively stable kernel releases that don't change substantially or that happens to be supported by ZoL with binary packages. ZoL on Arch was... trying at times. It worked great, but my paranoia meant that I ended up adding the kernel to IgnorePkg to force manual kernel updates (mostly for my o…

Post by mshook » Thu, Mar 23, 2017, 2:08 PM UTC

aarmenaa wrote: post #69

kzrdude wrote: post #30

Earlier quoted context omitted.

I think Chris Mason is open with that raid support is not really stable yet, right? Status page says so: https://btrfs.wiki.kernel.org/index.php/Status

As far as I know they consider RAID0, 1, and 10 to be stable. Last time I used it rebuilds were substantially slower than ZFS or mdraid. Rebuild performance seems to one of a few issues that BTRFS has had trouble solving. RAID 5 and 6 were declared stable last year, only to have that retracted when some fatal flaw was discovered that would apparently cause data loss if you needed to rebuild.