It's quite the opposite in fact. BTRFS was pulled INTO the kernel recently. It's seeing massive development and is production-ready.
> Why not ZFS?
The main reason for me is its flexibility. Just an example:
Expanding BTRFS is just a matter of shoving in arbitrary disks and rebalancing the existing array. ZFS requires you to plug in drives of the same size to properly scale.
I suppose Ive been using ZFS in production for so long and its been so stable that anything that isnt fully distributed wouldnt be a step up for us. CentOS8 had ZFS in the kernel, I think debian still does (ubuntu file servers using AD and SMB does not work well in my experience tho), so it looks like im going to alma or rocky. I guess Ill figure out which file system we will use at that point depending on native support. There are other ways to scale ZFS too but I have also found trying to mix and match drives, machines in a production environment is going to bring its own issues.
