Btrfs as default filesystem for Fedora 23?

[Chris Murphy]

On Tue, Jun 23, 2015 at 11:41 AM, Josh Boyer <jwboyer at fedoraproject.org> wrote:
> On Tue, Jun 23, 2015 at 1:30 PM, Andre Robatino
> <robatino at fedoraproject.org> wrote:
>> Neal Gompa <ngompa13 <at> gmail.com> writes:
>>
>>> As I recall, Josef Bacik mentioned that he'd be pushing for Btrfs becoming
>> the default in Fedora 23. At this point, I'm personally convinced that it is
>> certainly ready and doable for F23.
>>>
>>> Perhaps other guys with more experience on this stuff could chime in with
>> feedback/information/etc, but it feels like we should start the process to
>> get everything ready for Btrfs being default in Fedora 23.
>>
>> I asked about this recently on #fedora-devel (I was the one who asked
>> originally on this list) and was told there are no plans to make it the
>> default yet. It's amazing that it was originally planned to be the default
>> on F16 (see https://fedoraproject.org/wiki/Btrfs ). But I don't want to see
>
> Someone created a wiki page proposing that.  It was never actually
> planned to be the default.

It was a feature approved by FESCo to make it the default for Fedora
16 instead of ext4.

http://meetbot.fedoraproject.org/fedora-meeting/2011-06-08/fesco.2011-06-08-17.30.log.html

But this came with some criteria to be met before freeze, namely
"btrfsck", which were not met and two months later Josef said it would
not be the default. Since he subsequently left Red Hat, only Eric
Sandeen has much Btrfs knowledge, but he works mainly on the RHEL
kernel and doesn't have time to help maintain Btrfs stuff for the very
new kernels Fedora uses beyond what upstream does.

And for that matter, no one upstream intends for serious regressions
to happen in Btrfs, yet they can and do happen. So the catch-22 with
Fedora kernels being so new, is anyone using Btrfs is going to be
among the first users to experience bug fixes, feature enhancements as
well as regressions. I don't know that having an experienced Btrfs
kernel developer on the Fedora kernel team would matter that much in
preventing regressions from landing as Fedora stable kernels. Rather,
it'd probably take an increase in time to stable or increase in karma
value to delay the unknown, until better known.

openSUSE uses Btrfs by default (except for /home) but they'd also
using much older kernels with cherry picked backported bug fixes. Even
if Fedora had a Btrfs dev I don't foresee Fedora running 3-4 major
versions of the kernel behind, just to make Btrfs the default.

-- 
Chris Murphy