my dual CPU 2.0 machine ran out of descriptors last night during the
/etc/daily run:

file: table is full - increase kern.maxfiles or MAXFILES

this happened once before, and I was never able to track it down
definitively before I rebooted the machine.  I figured it had something
to do with me ripping CDs on an old IDE drive which was throwing up IDE
errors like crazy, and there being an obscure leak somewhere.  but as
that drive has been removed, I'm left with SCSI only, so it wasn't an
issue this time.

since this happened again, I'd like to diagnose the problem.  the man
pages indicate "pstat -T" should give me what I'm looking for:

$ pstat -T
1735/1772 files
  30663 vnodes
0/2098295 blocks swap space

that's pretty close to the edge.  my other machines run nowhere near
that kind of file load; even my busy mail server is 391/1772 .

so how to diagnose this?  "fstat"?

$ fstat -n | wc -l
     845

if fstat isn't reporting all the open descriptors on the machine, how
else can I diagnose this?

-- 
  Aaron J. Grier | "Not your ordinary poofy goof." | agrier@poofygoof.com
    The United States is the one true country.  The US is just.  The US
    is fair.  The US respects its citizens.  The US loves you.  We have
    always been at war against terrorism.