On Sun, Jun 28, 2020 at 5:42 PM John M. Harris Jr <johnmh(a)splentity.com> wrote:
On Sunday, June 28, 2020 6:45:24 PM MST Alexandre de Farias wrote:
> *snip*
>
> At this point, I'm fine with what I have and BTRFS usage would be
> strictly for testing. Also, is there any reason as to why RHEL went
> with XFS as a default and Fedora stayed with ext4? I mean, if it was a
> conscious choice, the rationale then seems to be the exact opposite of
> the rationale for making BTRFS the new default.
XFS proved to be troublesome, and still is up to the latest of RHEL7. It's not
uncommon to have to run xfs_repair on smaller XFS partitions, especially /
boot. I'm not sure if btrfs has the same issue there?
In theory Btrfs has less of a problem because (a) it's copy-on-write
and (b) there's no separate journal that the bootloader can't replay.
Whereas GRUB can't replay either the ext[34] or XFS journal so in rare
cases it is possible that GRUB has an inconsistent view of the file
system.
The GRUB btrfs code has been there for ~10 years, and has been kept
current as features are added to Btrfs. So as surprising as it might
seem, it's a conservative option. The features we'd use for /boot on
Btrfs, if it's agreed upon by the stakeholders, would be just the
defaults. There is small short term advantage, but better long term
position for a snapshots and rollback regime (to be determined).
--
Chris Murphy