Some named things I've learned ...

-----------------------------------------------------------------------------------
  In AIX 4, IBM ships both named versions 4 & 8.  By default, it runs version 4.

/usr/sbin/named      -> /usr/sbin/named4
/usr/sbin/named-xfer -> /usr/sbin/named4-xfer

  You can select which you want to run by changing those last two links.
If you do change them, change both and keep them in synch.  For example, on loon
I switched to using version 8 back in August 1, 2001 by changing the links to

/usr/sbin/named      -> named8
/usr/sbin/named-xfer -> named8-xfer

-----------------------------------------------------------------------------------

By default, named version 4 reads /etc/named.boot, version 8 reads /etc/named.conf.

-----------------------------------------------------------------------------------

  One debugging technique, is to send a kill -INT signal to the named PID, which
causes the daemon to dump its internal database to /var/tmp/named_dump.db.

  Another technique is to send it a -USR1 signal to increase its debugging level.
This ongoing, run-time trace is written to /var/tmp/named.run.
Send it a -USR2 signal to shut tracing off.

  And of course, you've always got the syslogd logging you can look at.
Check your /etc/syslog.conf file.

-----------------------------------------------------------------------------------

  Another debugging technique is to start named with startsrc -s named -a "-d 4",
and check the /var/tmp/named.run file.  This grows quite quickly and frankly, I
don't understand much of that file.

-----------------------------------------------------------------------------------

  I was looking into why loon wasn't doing zone transfers to tiger or chitown
and came across this note from the http://www.isc.org/products/BIND/bind9.html
file,
        A bug in the Windows 2000 DNS server can cause zone transfers from a
     BIND 9 server to a W2K server to fail.  For details, see the "Zone Transfers"
     section in doc/misc/migration.

so I downloaded the latest bind code from http://www.isc.org/products/BIND
on 8-21-2002, unzip'd it via
  cd /afs/d/software/base
  gzip -d < ~/bind-9.2.1.tar.gz  | tar xf -
and looked at their doc/misc/migration file, which said
        Zone transfers to Windows 2000 DNS servers sometimes fail due to a bug
     in the Windows 2000 DNS server where DNS messages larger than 16K are
     not handled properly.  There will be a hot fix available from
     Microsoft to address this issue.  In the meantime, the problem can
     be worked around by setting "transfer-format one-answer;".
     [As of May 4 2001 the hotfix was still being prepared]

In another section of that same file, it had this to say about
"transfer-format one-answer;"
        The default of the "transfer-format" option has changed from
     "one-answer" to "many-answers".  If you have slave servers that do not
     understand the many-answers zone transfer format (e.g., BIND 4.9.5 or
     older) you need to explicitly specify "transfer-format one-answer;" in
     either the options block or a server statement.

I stuck a "transfer-format one-answer;" line in the options block of loon's
/etc/named.conf file and that fixed tiger.

-----------------------------------------------------------------------------------

In Feb, 2006, Mike had troubles with name resolution on dweb12 (formerly dncdb2,
which had HACMP on it, which as it turns out, was relevant).  IP names wouldn't
resolve when under fcgi, but they did from the command line.  Turns out it was a
   hosts = local
line in /etc/netsvc.conf.  This evidently was a carryover from the HACMP config.
Mike took that line out and left it with just these lines, which are the EDC
default.
   netgroup=nis
   passwd=local, nis
   group=local, nis

The AIX default seem to be
   hosts=local ,bind4
   export NSORDER=local ,bind4