[Going back on-list, as I originally forgot to reply-all, and this might interest others.]

On Thursday, March 17, 2011 08:56:45 am you wrote:
> On Thu, 17 Mar 2011, Lamar Owen wrote:
> > Well, this is one that, due to the new kernel source packaging,
> > you'd have to ask upstream, since SL's kernel is a direct rebuild.
 
>   i suspected as much, but am i at least sane to think that this is a
> strange combination of packaging?

Yes, you're quite sane, assuming that btrfs can't possibly be supported by a 32-bit kernel.  After all, a 32-bit kernel can support (not well, but that's another matter) an XFS filesystem up to 16TB of occupied data; I got bit by that, with a 20TB LVM logical volume formatted XFS, with a 32-bit kernel.  When the amount of data on that filesystem went above 16TB (which it did), the kernel lost its ability to mount that filesystem.  MD5 sum checks on the files in that filesystem, once I restored access to it by installing a 64-bit CentOS, passed and so I knew no data was corrupted.

And, in fact, googling around I find in
https://bugs.launchpad.net/ubuntu/+source/dpkg/+bug/607632
that at least on Ubuntu a 32-bit kernel can do btrfs.

I don't see a mention of btrfs in Red Hat's release notes.  

It's listed as experimental in the Fedora documentation from Fedora 11; just checking now on my Fedora 14 32-bit box, I see:
[root@localhost ~]# yum list|grep btrfs
btrfs-progs.i686                        0.19-12.fc14                 fedora     

So F14 has 32-bit btrfs-progs.

Ok, so looking in the spec file for SL6/RHEL6:
%changelog
* Mon Jun  7 2010 Josef Bacik <[log in to unmask]> 0.19-11
- fix btrfs-progs so it only buils on x86_64, Resolves: rhbz#596690

Hmm, let's see:
https://bugzilla.redhat.com/show_bug.cgi?id=596690
Title: missing package btrfs-progs for i686
"BTRFS is going to be Tech Preview only for x86_64, you will not be able to use it on any other architecture."

Doesn't say why.