The cool thing here is that, although they don't have the posix layer done yet, they do have all the pooling stuff working. That gives you scrubbing, snapshots, RAID-Z{,2,3}, etc. In their example page (http://wiki.github.com/behlendorf/zfs/example-zvol), they show how you can make a zvol and put ext2 on it, which is a pretty cool step towards having reliable data on linux. Too bad about the license, but I'd certainly consider it for a home server.
Well, BTRFS[1], comparable to ZFS, is showing some success. I remember reading somewhere Linus even used it on his laptop. So one question would be if this is really that interesting? It's pretty likely that the Linux ZFS implementation will just lag behind the BTRFS effort, which is GPL.
Well, btrfs is a new filesystem with a completely new implementation, and it is still not in "production quality". It's in the mainline kernel, it works (I recently converted my / to btrfs), but still I had to check an option labeled experimental with a large warning while configuring the kernel.
On the other hand, this is just a port of actually working and stable FS already used in production with Solaris (and mabye FreeBSD?). Btrfs won't be ready until next few releases.
We are talking about the Linux port, which isn't really the same functionality that's already running on Solaris/FreeBSD. I've been running ZFS on FreeBSD for a while and it works pretty ok, but even that isn't the same as the functionality on Solaris.
The FAQ even clearly states that they only support ZVOL operations. Kinda like mdadm and lvm2, but without POSIX FS compability. Anyhow, my point is that this isn't "production quality" either.
But you are right of course in that the on disk format for BTRFS hasn't stabilized yet, last time I checked, and that it is a project in its childhood.
With Oracle slashing it's Sun acquisition into profitability and the fact that the licensing issues still surround ZFS and their apparent doubling down on the btrfs effort it may not be ready for a few releases but in the longer term it definitely shows more promise especially if they end up slowing zfs work. ZFS is certainly an amazing and well written fs, but its starting to look like its future is a little uncertain.
From your link: Btrfs, when complete, is expected to offer a feature set comparable to Sun's ZFS.
It's as silly as the politicians in my country. About three or four years ago, they paid an absurd sum of money for a super computer that would be the third of the world when complete. Big surprise: three years later, once completed, the machine does not make it even into the list of the top 100.
> In my opinion, the basic architecture of btrfs is more
> suitable to storage than that of ZFS. One of the major
> problems with the ZFS approach - "slabs" of blocks of a
> particular size - is fragmentation. Each object can contain
> blocks of only one size, and each slab can only contain
> blocks of one size. You can easily end up with, for example,
> a file of 64K blocks that needs to grow one more block, but
> no 64K blocks are available, even if the file system is full
> off nearly empty slabs of 512 byte blocks, 4K blocks, 128K
> blocks, etc. To solve this problem, we (the ZFS developers)
> invented ways to create big blocks out of little blocks
> ("gang blocks") and other unpleasant workarounds. In our
> defense, at the time btrees and extents seemed fundamentally
> incompatible with copy-on-write, and the virtual memory
> metaphor served us well in many other respects.
Describing btrfs and zfs as 'exactly the same' just because most of the features that are talking about or that you care about are the same isn't the whole truth. [Not directed at you, but most discussions seem to assume some sort of equivalency between the two filesystems.]
> The CDDL does not restrict modification and release of the ZFS source code which is publicly available as part of OpenSolaris. The ZFS code can be modified to build as a CDDL licensed kernel module which is not distributed as part of the Linux kernel. This makes a Native ZFS on Linux implementation possible if you are willing to download and build it yourself.
