>> [...].  So I ran df.  I still had a pre-rm df output on my screen,
>> and the difference in free space before and after was on the order
>> of a few tens of megabytes, nowhere near .7 terabytes.

> This indeed seems broken, not that I can tell where the problem could
> be.

As I wrote later, it appears to be related to indirect blocks.  I
suspect they're not getting written to disk, or not getting written
correctly somehow, or some such.

> I am curious though, was this done with soft dependencies enabled?

No.  The filesystem was 8k/64k, too, and at least one person wrote me
off-list saying "ISTR hearing of issues with [bsize=64k], but I have
not looked into it at all".  I've rebuilt the filesystem as 1k/8k and
will be pounding on it that way a bit.  If that fails our next try will
probably be with FreeBSD - much as I may dislike it, this is a "needs
to work" production machine, not one I can casually experiment with.
(Given the price of disk these days, I may invest in a handful of big
drives and set something up myself.)

> I unfortunately don't recall if you're now trying with 3.0+ or
> -current,

2.0, plus a patch from mycroft (quite some time back) to fix a problem
when exceeding 1T.

> It would definitely be interesting to know if this also occurs with a
> test -current system.

I tried 3.0 for >2T and it failed my tests.  I haven't tried it for the
sub-2T stuff.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B