[4024] in SIPB bug reports
New nightly_jobs.pl script installed on charon
daemon@ATHENA.MIT.EDU (Jonathan I. Kamens)
Mon Aug 9 13:50:52 1993
Date: Mon, 9 Aug 1993 13:49:35 -0400
From: "Jonathan I. Kamens" <jik@gza.com>
To: charon-maintainers@athena.mit.edu
Cc: ca-maintainers@athena.mit.edu, owner-mail-server@charon.MIT.EDU
Cc: jik@athena.mit.edu, bug-sipb@athena.mit.edu, root@charon.MIT.EDU
I have installed on charon a script in /etc/nightly_jobs.pl (with the
source in /afs/sipb/project/dec-charon) and the configuration file
/etc/nightly_jobs.conf. I have modified crontab by removing from it
the jobs that will be run by nightly_jobs.pl and by adding a line to
it to run nightly_jobs.pl at midnight. I have also created an RCS
directory in /afs/sipb/project/dec-charon/etc, and RCS'd crontab,
nightly_jobs.conf and nightly_jobs.pl.
Please read the nightly_jobs.conf file to find out what its format is
(there's a comment at the top which explains it). Basically, the
purpose of nightly_jobs.pl is to run a series of jobs sequentially,
capture their output and any non-zero exit statuses, and report them
to the job owners.
The jobs currently being run by nightly_jobs.pl, with the indicated
owners, are:
1) The Central America run (first, since people want to see the day's
plans early in the day and it takes a long time to run). (ca-maintainers)
2) The nightly mail-server lock cleanup. (owner-mail-server)
3) The nightly mail-server collection of votes from the server logs.
(jik)
4) The nightly FAQ archiver (rkive) run. (jik)
5) The nightly "List of Periodic Informational Postings" run. (jik)
6) The rebuild of the FAQ WAIS index (although this only actually
happens on Sunday). (jik)
7) The rebuild of FILES and FILES.look in sipbsrc. (bug-sipb)
8) The mail-lossage script which checks for local mail on charon. (root)
9) The check-backups and run-clean-db scripts for the backup system.
(root)
I didn't move "/etc/athena/newsyslog -f /usr/local/nje/newsyslog.conf"
because I don't know whether Matt wants it to keep running as close to
midnight as possible. If that's not important, it should probably be
moved too (although granted, it's not an intensive enough job that we
need to worry about it).
I've tested the script by running the jobs that can be run safely more
than once in a day, and all of them worked fine. The script should
log any fatal errors to daemon.err and also send mail to root about
them.
Comments?
jik
