TRACK INSTALLATION INSTRUCTIONS:

0) choose a librarian machine and unbundle the track source files
   there. (you have probably done this already).

1) Choose a working directory for track, and change DEF_ROOT in
   src/Makefile.  Modify any other variables in that Makefile to your
   taste.

2) If you really need to, you can modify the definitions in src/track.h 
	DEF_ROOT	track's working directory
			(this can be changed at runtime
				by using the -d and -r flags)
	DEF_LIB		the default librarian
			(this can be changed at runtime by
				using the -l flag)
	DEF_ADM		the person who receives mail if trouble
				occurs
			(this can be changed at runtime by
				giving an argument to the
				-m flag)
	DEF_SUB		default subscription list
			(this is usually changed at runtime by giving
				a final, unflagged argument to track)
	DEF_EXCEPT	default exception list
			(these exceptions are added to all entries
				at runtime and can't be bypassed)
	other defaults may be changed, but only at great risk and
	little benefit.

3) run make in the src directory.

4) install track on the librarian with 'make install'.

5) If you want, create a permissions file on the librarian
   (see permissions(5)) and put your subscriber machine names in it.
   If you do not create a permissions file, anyone can track any
   slist from your system.

7) create an empty file named logfiles/lib.log if you want to log
   activity on the part of the librarian.

8) Add an entry of the form

track           3462/tcp        track/shell

   to your /etc/services files.

tracktest	4939/tcp			#special track test port

   as well, if you want to compile track with -DDEBUG (See src/Makefile)

9) If you want detailed logs, run the librarian as
	trackd -d
   in which case it will log to the standard error and not
   background itself as a daemon.  The first time, it is probably
   best to run track like this, with stdout/stderr directed at a file.

   Otherwise, start trackd as 
	trackd
   and it will log errors to syslog.  In this case, add the lines to your
   /etc/rc.local or /etc/init.d/local to start trackd when your
   librarian reboots.

   Alternatively, you can make an entry of the form

track   stream  tcp     nowait  root  /local/lib/track/bin/trackd trackd -inetd

   in your inetd.conf file and have trackd started from inetd.


10)Install track on your subscribers.  If they are the same architecture
   as your server, you can copy/NFS-mount the source directory
   and type 'make install' on the subscriber as well.  Otherwise,
   you need to recompile the source there.

11)Create a simple slist on the librarian and the subscriber in the
   slists directory,
   (eg.)

/tmp/somefile::::

   Generate the statsfile for the simple slist on the librarian as follows:

	track -w simple

   Then try track on the subscriber as:

	track -l librarian simple

   (If you want, you can add the -v or -n options to see what
   it does, or is going to do)

   The client should drag over the simple statsfile from the server,
   compare it with it's own files, and bring /tmp/somefile across.
   Experiment with the slist syntax if you want.  The slists in
   the slists/ subdirectory may give you some ideas.

11)Once you've got track working, create your slists.  Choose
   the times you want the librarian to generate the statsfiles for
   those slists.  Then add entries to your crontab to run "track -w"
   for each slist at the chosen times.

   Then you can choose the times and make up the crontab entries
   for the subscribers to check with the librarian.  There's no point
   in a client running track on an slist till the librarian has run
   track -w on that slist.  It's also best to have the clients avoid
   bashing on the server at the same time.

Enjoy.
