On Wed, 11 Mar 1992 14:36:55 EST Rex Bontrager <REXB@PURCCVM> said:
 
>We (Purdue University) are a backbone LISTSERV, but our downstream sites
>are not.
 
Write them hate mail every day until they accept to join the backbone :-)
 
>Why are all LISTSERV sites not setup to be backbone sites?
 
Say you  were to  shut down  your LISTSERV  for 2  weeks because  of some
important local  event requiring  the full  attention of  your CPU  for 2
weeks,  like fiscal-year-end  administrative  things with  a deadline  or
whatever. Since you are a backbone server, the people downstream you will
get no LISTSERV mail for 2 weeks; I doubt they will be happy.
 
>So, please  enlighten me with the  pros and cons of  turning most sites,
>all sites, or "all sites which diligently keep their versions of BITEARN
>NODES current" into LISTSERV backbone sites.
 
There is no con  as long as you stick to sites which  treat LISTSERV as a
service, and not an additional goodie they start up as the need arises. I
have execs to review the status  of backlevel servers, many sites tell me
LISTSERV NOT  LOGGED ON,  or occasionally LISTSERV  NOT IN  CP DIRECTORY.
Many run  VERS8811, VERS9001, or  the like. You  don't want these  on the
backbone.
 
>Another question  from my boss:  Could LISTSERV  be modified so  that it
>would automatically apply updates to BITEARN NODES,
 
No, because it doesn't *get* updates to BITEARN NODES in the first place.
There is a NODEUPD  PACKAGE (not updated for the new  format, not sure it
still works) you  can install to make LISTSERV run  UPDNODES for you, but
it still requires you to subscribe it to the files, take care of renewing
the subscription every year, and so on. You need NETSERV authorization to
subscribe  to   files.  There  are  also   virtual  storage  requirements
associated with  this function, and  obviously you  need a bunch  of disk
space to  keep 3 copies of  BITEARN NODES (current, last  month's in case
the current is broken  and you have to back out, and  the one being built
by UPDNODES).
 
>or simply shutdown if its BITEARN NODES  file was more that n months out
>of date,  where n is hardwired  and not a parameter?  I'm not advocating
>doing this (yet), but the shutdown  part of the question was amusing and
>I would like to hear other people's comments on the concept.
 
That would increase  the amount of LISTSERV NOT LOGGED  ON sites. Believe
me, that  will NOT  make local staff  take action. In  the case  of sites
running a very  old version of BITEARN NODES, local  staff either doesn't
know there is such a thing as  LISTSERV on their machine, or doesn't give
a damn. I have  a program I run each month to remind  people to send me a
PEERS NAMES entry. Some have been  getting mail from me every month since
1988,  and never  bothered  to reply.  One of  these  sites noticed  that
LISTSERV no longer worked; instead of thinking about fixing it, they just
wrote an  EXEC to purge the  files in its spool  once a day (I  found out
after a staff change). Do you want these on the backbone? :-)
 
  Eric