Btrfs Enjoys Performance Optimizations With Linux 6.9

[phoronix]

Phoronix: Btrfs Enjoys Performance Optimizations With Linux 6.9

SUSE's David Sterba sent out the Btrfs updates today for the Linux 6.9 merge window. Besides stabilization and bug fixes there are also some minor performance optimizations to see with this next kernel...

https://www.phoronix.com/news/Btrfs-Linux-6.9

[S.Pam]

Always nice with improvements. 6% increase in throughput seems more than a minor increase

[stiiixy]

Bug fixes and incremental speed improvements means a happy BTRFS user. Big changes to a filesystem to me are scarey 😱

[EphemeralEft]

Nice.

[GreenReaper]

Always nice with improvements. 6% increase in throughput seems more than a minor increase

Depends on where the bottleneck is. I have seen major improvements to scrub CPU usage in the past, resulting in a speedup. But anything that is based on having fast storage will have no impact because I use it on HDDs.

[starkruzr]

is Btrfs still a serious project? I feel like it has been so completely eclipsed by ZFS and don't really know what it even offers over XFS or ext4.

[oleid]

is Btrfs still a serious project? I feel like it has been so completely eclipsed by ZFS

No, it is not. These are the final patches before they are going to remove btrfs completely from the kernel and change the kernel licence to be able to pull in ZFS.

and don't really know what it even offers over XFS or ext4.

That's easy: basically many of the features zfs provides over xfs and ext4. But I'm sure your already knew that and merely intended to provoke people.

[hyperchaotic]

is Btrfs still a serious project? I feel like it has been so completely eclipsed by ZFS and don't really know what it even offers over XFS or ext4.

Well, you can't compare filesystems on a feel. Especially if you don't understand the difference in capabilities between Cow and non-CoW filesystems. As for ZFS vs. BTRFS there's some overlap but also use cases where each fit better than the other.

[Britoid]

It would be nicer to see btrfs gain some optimisations for running a vm/sql db on it, rather than having to disable CoW.

[varikonniemi]

It would be nicer to see btrfs gain some optimisations for running a vm/sql db on it, rather than having to disable CoW.

Do you have a theoretical understanding of how to do this, even if it is not practically implementable? For me it's hard to see how COW could be made to compete with update-in-place.