Matthew Ahrens writes:
> Since we're exposing the both zfs and zpool "version" as a property, it 
> makes sense to be able to change it with "zfs set".  Since we've established 
> the precedent of "zpool upgrade" (and it has additional capabilities, eg. 
> "zpool upgrade -a" and "zfs upgrade -r"), it makes sense to have that as well.

OK.  It must just be me, then.  I find the idea of "set version" to be
a bit baffling.  I'd expect that (if it did anything), setting the
version would be forcing the system to lie about the current version
number, rather than changing the on-disk version.

It's sort of like seeing a variant on sysinfo(2) that can write to
SI_RELEASE.  I wouldn't expect that writing a number into there would
cause the system to be upgraded.

> Granted, if we could redo the interface from scratch, it would probably be 
> better to only change the version with "zfs/zpool set version"[*], and only 
> get version information with "zfs/zpool upgrade" (or some more appropriately 
> named subcommand).  However, given that we can't change the existing "zpool 
> upgrade" interface, this seems like the most self-consistent choice.

I see.  I guess it doesn't bother me enough to complain more than
that.

-- 
James Carlson, Solaris Networking              <james.d.carlson at sun.com>
Sun Microsystems / 1 Network Drive         71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677

Reply via email to