Re: Redhat dropping support for BTRFS

[Nathaniel Reindl <nrr@corvidae.org>]

On Sat, Aug 12, 2017, at 21:11, Steven Pritchard wrote:
> So I'll admit to being lazy and not looking at the PDF you linked to,

Laziness is a good virtue to have here. (:

> but I would be willing to bet that Stratis doesn't solve the major
> problem that ZFS and btrfs solve: detecting bit rot and other sources of
> corrupt data on disk.  I understand there is some work being done on

Yeah, so, the document I linked actually speaks to that in §10.2, noting that they're intending on using some machinery layered atop device-mapper to make it all work. The specifics are lacking, but I suspect that's because this is literally, like, less than a week old.

> adding checksum support to device-mapper, which could conceivably solve
> the problem for all users, but I don't know how close that is to being
> readily available.

Right, and that document also helpfully states "in a later version" in the footnotes when it comes to data integrity. However, being built atop XFS, there's some degree of data integrity detection and repair built in by way of xfs_repair, so not all is lost out of the gate, at least as far as the on-disk representation is concerned.

How Stratis intends to integrate that into something that's ZFS-alike, well, we'll see, I guess.

> Yeah, I agree...  I'm not entirely convinced that adding the packages
> from zfsonlinux.org to RHEL would be any kind of a licensing issue, but
> I am not a lawyer...

Nor am I, but I thought I'd toss that out there just to illustrate just how ridiculous I find this whole situation. —n


-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.