sidai:dchunk-utils/acct-for-diskspace

Just for sport... here's one I kinda like...

On the glasli1 team, we're experimenting with 'disk chunks'
(dchunks); if you look in the team stuff we distrib, in its
'chunk' dir, you'll see one xml file per
interesting-chunk-of-disk, be it a partition or a thing that
goes out in an automount map...

If you describe *all* of a site's disk space this way, then
you can generate your /etc/exports, your automount maps,
your Amanda disklist (backups manifest), and some other
stuff not yet thought of.

You can also then go searching for cruft on your disks.
That's what 'acct-for-diskspace' does; witness the [slightly
edited] output below.

Now, never mind the output format... the point is that
everything it coughs up is (a) either junk or (b) needs to
be accounted for in the dchunks' info.  I sleep better at
night when this kinda stuff gets bolted down.

Site-at-a-time: one command invocation does the whole site.

Will

% PYTHONPATH=/our/share/ark ./acct-for-diskspace
ignoring non-active host: clarence
R slinger:/._disc1/._our-deploy
- slinger:/._disc1/._our-deploy/PyXML-0.5.5.1
- slinger:/._disc1/._our-deploy/Tk-800.015
- slinger:/._disc1/._our-deploy/a2ps-4.12
R slinger:/._disc1/OLD
- slinger:/._disc1/OLD/._ark
- slinger:/._disc1/OLD/._ark-deploy
- slinger:/._disc1/OLD/._our
R slinger:/._disc1/item/g141
- slinger:/._disc1/item/g141/cvsroot
R slinger:/._disc1/item/matt-ark-src
- slinger:/._disc1/item/matt-ark-src/ARK
- slinger:/._disc1/item/matt-ark-src/glasli1
- slinger:/._disc1/item/matt-ark-src/matt1
- slinger:/._disc1/item/matt-ark-src/sidai
R slinger:/._disc1/scratch
- slinger:/._disc1/scratch/matt
R slinger:/._disc2
- slinger:/._disc2/lost+found
- slinger:/._disc2/test-recover
R slicker:/._disc1
- slicker:/._disc1/lost+found