For Limerick, who's running the new IBM Network Dispatcher code,
the commands are
  dscontrol manager report (10.225.49.11)
  dscontrol server down 10.225.49.11:3700:ddb2
  dscontrol cluster set 10.225.49.11 staletimeout 2592000 (also .12-.15)
  dscontrol port status www2x:443 

IBM's "Load Balancer Administration Guide" is online at
http://www-306.ibm.com/software/webservers/appserv/doc/v51/ec/infocenter/edge/LBguide.htm

The Network Dispatcher is now packaged in filesets starting with ibmlb,
which evidently stands for IBM Load Balancer.  The code resides in
/opt/ibm/edge/lb and principally in, /opt/ibm/edge/lb/servers/bin.


The alias commands need to be
  On the ND: ifconfig en0 delete 10.225.49.11 
             ifconfig en2 alias  10.225.49.11 netmask 255.255.252.0
             ifconfig en0 alias  172.16.49.11 netmask 255.255.252.0
  On each server: ifconfig lo0 delete 10.225.49.11 
                  ifconfig lo0 alias  10.225.49.11 netmask 255.255.255.255
                  ifconfig lo0 alias  172.16.49.11 netmask 255.255.255.255

On some systems, I don't know for sure, you may need to delete routes
when deleting aliases, e.g.
  route delete 10.225.49.11 10.225.49.11
but I've never really come across when/where this bad route was there.
That's what the book says.

=======================================================================
        OO   L     DD        SS   TTTTT  U   U  FFFF  FFFF      Pre-Y2K
       O  O  L     D D      S       T    U   U  F     F
       O  O  L     D  D      SS     T    U   U  FFF   FFF
       O  O  L     D D         S    T     U U   F     F
        OO   LLLL  DD        SS     T      U    F     F


Some Network Dispatcher commands,

   ndcontrol cluster status 204.146.135.160
   ndcontrol server  down   204.146.135.160:80:192.168.56.16
   ndcontrol server  up     204.146.135.160:80:192.168.56.16

Additionally, there's
   ndcontrol server  delete 204.146.135.160:80:192.168.56.16
   ndcontrol server  add    204.146.135.160:80:192.168.56.16
You can do this to (probably) "fix" the high "Active:" count
problem when you do a ndcontrol cluster status.

-----------------------------------------------------------
Web pages are at .../nd.html or the online docs are at
http://www-4.ibm.com/software/webservers/perfpack/resourcesend.html

which showed the PDF is at
http://www-4.ibm.com/software/webservers/perfpack/assets/nduser.pdf

Page 63 of the book (pg 79 of the PDF file) is where it said
   If you are not using the Dispatcher's wide area support, a
   Dispatcher configuration requires that the Dispatcher
   machine and its servers all be attached to the same LAN
   segment.
Their "wide area support" refers to a ND dispatching to
another ND, then to servers.

-----------------------------------------------------------
There're ND log files in the /usr/lpp/eND/dispatcher/logs
directory, which I've got linked over to /var/adm/sng.

I've also got these lines in my /etc/eND/ind_setup_cluster
to control the log file,

  #  The old version of ND kept a log file in /tmp/manager.log, but
  #  this version defaults to /usr/lpp/eND/dispatcher/logs/manager.log,
  #  which can (and did) easily fill up poor /usr.  I moved it to the
  #  firewall logging, 1GB filesystem, and limit it's size to 1MB,
  #  keeping a backup copy as well.                  RAJ   10-15-1998
  mv /var/adm/sng/ND_manager.log /var/adm/sng/ND_manager.log.old
  ndcontrol manager start
  ndcontrol manager logsize 1048576

-----------------------------------------------------------
When the Network Dispatcher code takes a dump, look in the
/etc/eND directory for core files.  For example,
<root@as0176e1:/etc/eND> ls -ltr *core*
-rw-r--r--   1 root     security   17431 Mar 23 15:35 javacore.txt
-rw-r--r--   1 root     security 1048576 Mar 23 15:35 core