cronsched is killing my daemon? HLP PLZ

cronsched is killing my daemon? HLP PLZ

Post by David Smar » Thu, 03 Aug 2000 04:00:00



My SCO OpenServer 5.05 has had me running in circles this week - I
just hope that someone out there has a clue.

Problem: Most nights recently the nightly backup job has failed to
run - and - the Btrieve (Faircom CTree) server is also down in the
morning. When I finish killing and restarting the CTree server to
get the users application online, then the backup job starts right
on up. AARG
Then I have to do a recovery on the databases since the server
crashed with some users who forgot to logout.

And I swear "I haven't changed a thing in months!" <<right!>>

To debug - I turned on cron logging (in /etc/defaults), and set up a
NOHUP script to snapshot the "PS -ef" every 30 secs to catch some
system state before and after the server dissappears ... and lo ..
it failed right about midnight - when cron log and PS listing shows
it kicked off "cronsched -r".

---snipit from /usr/lib/cron/log

Quote:>  CMD: ulimit 5000; /usr/lib/uucp/uudemon.clean > /dev/null
>  uucp 27544 c Tue Aug  1 23:45:00 2000

<  uucp 27544 c Tue Aug  1 23:45:01 2000

Quote:>  CMD: scosh cronsched -r
>  root 27985 c Wed Aug  2 00:00:00 2000

<  root 27985 c Wed Aug  2 00:03:52 2000
 rc=1   <===== shows rc - and apparently killed the ctbg daemon!!

Quote:>  CMD: /usr/lib/uucp/uudemon.hour > /dev/null
>  uucp 28073 c Wed Aug  2 00:09:01 2000

<  uucp 28073 c Wed Aug  2 00:09:01 2000
---

What the blazes is cronsched ? And why might it be destructively
interfering with the server ??? And why now after years of faithful
service ?? Is this a sighting of the dreaded Y2K monster at last??

Any help, pointers to man pages, or websites is greatly appreciated.
This daily drill is wearing thin.  
Dave Smart

 
 
 

cronsched is killing my daemon? HLP PLZ

Post by David Smar » Sat, 05 Aug 2000 04:00:00


OK .. so maybe cronsched is not the culpret killing my daemon ... :)
 Since I removed the midnight cron entry .. no server crashes.
 Added cronsched back on cron but at 8PM .. no server crashes.

I ended up removing a lot of junk started from rcd2.d
(rcpinfo, calander, faststart) and things seem much more stable
this week. And installed a new chassis fan.

Dave


> My SCO OpenServer 5.05 has had me running in circles this week - I
> just hope that someone out there has a clue.

> Problem: Most nights recently the nightly backup job has failed to
> run - and - the Btrieve (Faircom CTree) server is also down in the
> morning. When I finish killing and restarting the CTree server to
> get the users application online, then the backup job starts right
> on up. AARG
> Then I have to do a recovery on the databases since the server
> crashed with some users who forgot to logout.

> And I swear "I haven't changed a thing in months!" <<right!>>

> To debug - I turned on cron logging (in /etc/defaults), and set up a
> NOHUP script to snapshot the "PS -ef" every 30 secs to catch some
> system state before and after the server dissappears ... and lo ..
> it failed right about midnight - when cron log and PS listing shows
> it kicked off "cronsched -r".

> ---snipit from /usr/lib/cron/log
> >  CMD: ulimit 5000; /usr/lib/uucp/uudemon.clean > /dev/null
> >  uucp 27544 c Tue Aug  1 23:45:00 2000
> <  uucp 27544 c Tue Aug  1 23:45:01 2000

> >  CMD: scosh cronsched -r
> >  root 27985 c Wed Aug  2 00:00:00 2000
> <  root 27985 c Wed Aug  2 00:03:52 2000
>  rc=1   <===== shows rc - and apparently killed the ctbg daemon!!

> >  CMD: /usr/lib/uucp/uudemon.hour > /dev/null
> >  uucp 28073 c Wed Aug  2 00:09:01 2000
> <  uucp 28073 c Wed Aug  2 00:09:01 2000
> ---

> What the blazes is cronsched ? And why might it be destructively
> interfering with the server ??? And why now after years of faithful
> service ?? Is this a sighting of the dreaded Y2K monster at last??

> Any help, pointers to man pages, or websites is greatly appreciated.
> This daily drill is wearing thin.
> Dave Smart