hi Sean,
On Wed, Oct 17, 2001 at 09:33:44PM -0400, Sean Elble wrote:
> So I should try to grab the SourceForge package instead? The SGI version is
> fairly current (pre-4 I think), but I'll give anything a shot at this point.
> I enabled XFS quotas _and_ the other quota thing in the kernel . . . here is
> a snippet of my start-up:
Aha, very useful information.
> SGI XFS with ACLs, EAs, DMAPI, quota, no debug enabled
> [ . . . ]
> XFS mounting filesystem ide0(3,5)
> XFS quotacheck ide0(3,5): Please wait.
> XFS quotacheck ide0(3,5): Done.
>
> It looks like everything is working correctly, until I try to do quota on.
Cool - so quota are on already! What was the problem again? ;-)
You don't need to use quotaon(8) for XFS quota here - XFS quota are
enabled at mount time. This is one of the areas where XFS quota
differs quite a bit to other implementations of quota. In XFS the
quotaon, quotacheck, and mount operations are one operation - the
time window between these operations which exists for other quota
implementations (wherein it is possible to write to the filesystem
before quota are enabled) has been removed, so it is impossible to
"corrupt" the quota accounting [meta]data in this way.
Check out README.quota... it'll fill in the missing pieces for you.
> :-( I think the SGI quota tools isn't "special", but you just provide it as
> a "nice thing" . . . you know better than I do though. :-) Thanks again, and
> have a good one; I'll try to build tommorow.
You don't need to rebuild - from the XFS kernel initialization line
above we can see that quotas have indeed been compiled in.
cheers.
--
Nathan
|