[comp.sys.pyramid] Disappearing cron

law@ioe.lon.ac.uk (Lindsay Wakeman) (11/07/90)

We have a 98xe running 4.4. From time to time (soon after a reboot)
we lose our ucbcron daemon. rc.local starts both att and ucb daemons.

Anyone else seen this or know a reason? 
Thanks
Lindsay
-- 
JANET: law@uk.ac.lon.ioe			| Lindsay Wakeman
EARN/BITNET: law%ioe.lon.ac.uk@ukacrl.bitnet	| Institute of Education
INTERNET:law%ioe.lon.ac.uk@nsfnet-relay.ac.uk	| 20 Bedford Way London WC1H OAL
UUCP: !mcvax!ukc!educ-isis!law			| VOICE +44 71 636 1500 ext.512

law@ioe.lon.ac.uk (Lindsay Wakeman) (11/07/90)

In <1990Nov6.165642.2742@ioe.lon.ac.uk>  I wrote:


>We have a 98xe running 4.4. From time to time (soon after a reboot)
                      ^^^^^^^^

Sorry, complete abberation here (or wishful thinking?) - We're actually
running OSx 4.1

-- 
JANET: law@uk.ac.lon.ioe			| Lindsay Wakeman
EARN/BITNET: law%ioe.lon.ac.uk@ukacrl.bitnet	| Institute of Education
INTERNET:law%ioe.lon.ac.uk@nsfnet-relay.ac.uk	| 20 Bedford Way London WC1H OAL
UUCP: !mcvax!ukc!educ-isis!law			| VOICE +44 71 636 1500 ext.512

rpp@rilin.ci.OZ.AU (Richard Perini) (11/08/90)

In article <1990Nov6.165642.2742@ioe.lon.ac.uk> law@ioe.lon.ac.uk (Lindsay Wakeman) writes:
>
>We have a 98xe running 4.4. From time to time (soon after a reboot)
>we lose our ucbcron daemon. rc.local starts both att and ucb daemons.
>
>Anyone else seen this or know a reason? 

Yes!  We have seen exactly the same thing on an MIS-2 with OSx5.0.  
There is no core dump and no error message.  I haven't got round to 
reporting it yet...

-- 
Richard Perini                                             Phone: +61 2 9064333
Corinthian Engineering                                       Fax: +61 2 9061556
Sydney, Australia                                 Internet: rpp@sphinx.ci.oz.au

csg@pyramid.pyramid.com (Carl S. Gutekunst) (11/09/90)

The ucb cron in OSx 4.1 definitely disappeared on a regular basis. I remember
this all too well from our E-Mail/News machine getting backed up to rafters
when cron died yet again. I recall that it had something to do with cron not
protecting itself against signals, that is, cron and the tasks it started
ran in the same procress group, so that when a task signal itself or was sent
a signal, it got the shell, too. Of course, this is ancient history; so I may
be thinking of something entirely different. (OSx 4.1 has been outdated for
over three years. On a 98xe, there is no reason not to be running at least
OSx 5.0 or 5.1.)

Something to try: instead of running commands directly out of crontab, use a
C Shell script that calls the commands. C Shell always changes the process
group of commands that it runs, so if signals are the problem, then using an
intervening C Shell script should fix it.

<csg>