I got email working from the 3000 (5.5pp4, IIRC) yesterday,
but then took it back down.

The concern is that with three supporting jobs (one each for
syslog, bind, and sendmail), if any one of them crash, then
the job limit would allow user jobs to run simultaneously,
causing some to bomb.

My questions are:

1. has anyone seen any of these jobs fail (after having
   run successfully for "awhile") ?

2. is there a convenient way to bundle these together in
   a single job that does not fail ?  I'm thinking here
   of something like Allegro's HTTPBACK, or 'nohup' followed
   by an infinite 'while' loop containing only pause, or
   Vesoft's BACKG, or ....

(Multiple job queues" aren't bundled with FOS, are they?
Even if they are, I have no idea when this machine will be
updated beyond 5.5pp4.)

All tips are greatly appreciated.


BTW, installation was pretty painless, though configuring
BIND made me a bit nervous; I ended up commenting out most
of the zone-something config file.  Also, I deleted at least
one source tree, as the machine was running low on disk space.

On that particular machine, /etc/resolv.conf was an independent
file, i.e., NOT a link to RESLVCNF.NET.SYS.  IIRC, I had created
it myself a week ago based on instructions on the (a?) sendmail/iX
page.  I changed it to be a link, and modified RESLVCNF instead.

And just moments ago, I saw that the job for BIND does not need
to run.  I probably did not read the instructions closely enough.
Time to reread Alfredo's "love letter" paper. ;)

--Glenn

P.S.  I can see it now.  Sometime soon, there will be one of those
      "you're this old" spams making the rounds that includes the
      assertion, "You remember the Love Bug as a car named Herbie."