[14867] in Athena Bugs
sgi [8.0I]: install process
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Mon Dec 9 15:52:23 1996
Date: Mon, 9 Dec 1996 20:50:03 GMT
To: bugs@MIT.EDU
Cc: jweiss@MIT.EDU, sipb-staff@MIT.EDU, esg-cluster-mgrs@MIT.EDU
From: Jonathon Weiss <jweiss@MIT.EDU>
System name: deliverator.MIT.EDU
Type and version: IP22 8.0I
Display type: GR3-XZ
What were you trying to do?
See whether some of the Suns I help maintain currently had
mail locked in the mailq due to the sendmail lossage I've sent mail
about in the past, but that isn't really important. I happened to
check an sgi too, and that is where I found the real problem.
What went wrong?
I was surprised to find that the SGI had a nearly month old
piece of mail sitting unlocked in it's mailq. This is symptomatic of
the bug where the SGIs were not running their mailq's out of cron
every hour. We thought we'd fixed this bug in 8.0G, but we don't seem
to have gotten it completely. Based on my observations I believe that
newly installed SGIs are getting the wrong crontab. Specifically, I
looked at two SGIs, one installed on Fri Nov 15 12:35:41 PST 1996, and
the other installed in the last few weeks.
What should have happened?
SGIs should run the sendmail queue out of cron once an hour.
Yo, got any documentation, or other info?
The crontab that was installed was identical to
/install/var/spool/cron/crontabs/root, not
/var/spool/cron/crontabs/root.
sipb-staff: I ran crontab /srvd/var/spool/cron/crontabs/root on
steve-dallas, I didn't look at opus to see if it was suffering from
the same problem.
esg-cluster-mgrs: I ran crontab /srvd/var/spool/cron/crontabs/root on
torque.
