INN uses the DBZ database package.  Thanks to  Jon  Zeeff  for  his
permission  to  use  and redistribute DBZ, as modified by Henry Spencer.
INN has its own set of modifications to DBZ.  The changes are made  with
the  patch  program  and  the context diff in lib/dbz.pch.  If you don't
have patch installed, then you can make the changes manually.   (If  you
don't  have Larry Wall's patch program get it from any comp.sources.unix
archive as well as many FSF archives and other places -- you'll be  glad
you did.)

     Both innd and nnrpd have the option of keeping the DBZ  hash  table
in    memory,    under   the   control   of   the   INND_DBZINCORE   and
NNRP_DBZINCORE_DELAY parameters, respectively.  This can consume lots of
RAM  proportional  to the size of your history database, but it can also
avoid a great deal of disk I/O.  You should probably see the DBZ manpage
in  the  doc  directory  for  some (brief) additional discussion of this
issue.

     If you are using vfork (specified in the FORK  parameter),  or  you
want  to mmap the database, then you must apply the patch.  The Makefile
in lib will normally do it for you automatically, anyway.  The beginning
of  the patch file describes the changes made in more detail.  If you do
not apply the patch, then you must add add ``dbzalt.c'' and ``dbzalt.o''
to the MISSING_SRC and MISSING_OBJ parameters.

     Apparently the System V 386 compiler can't optimize dbz.c (the  GNU
C  compiler  doesn't  have  this  problem).   If you have ``-O'' in your
DBZCFLAGS configuration parameter, then take it out.