Re: [cobalt-users] Named update/Reboot/Logs dead/Get's me Red

[Jeff Lovell <jlovell@xxxxxxx>]

On Wed, 2001-11-14 at 06:26, David Thurman wrote:
> Hello All,
> We just installed the new Named update and this required a reboot of the
> server. This seems to be a constant problem with a system reboot. Our logs
> now do not rotate. I noticed that the clock moves forward 2 days. I have it
> set with a NTP server that is rock solid. I also noticed that the system and
> hardware clocks are now off. I resynced the clocks to each other but it
> would not take. I went in this morning and was able to resync the clocks to
> the correct time. We have 2 Raq3's and only one of them seems to act up. The
> one with the problems is a stock unit that was upgraded by Sun-Cobalt. The
> other with out issues was upgraded by us. I have had Tech support from sun
> in this machine before for the same reason. My question is, since I have the
> clocks resynced, will the cron now catch up and rotate the logs when it
> reaches Nov 16th (this is the day that was showing)

Yes.  Well, in reality it isn't cron that needs to catch up.  Cron isn't
that intelligent.  It just runs things when you tell it to.  Logrotate
is the intelligent piece here that has set the timestamp for the log
file in it's status database.  So it will run again after November 16th,
or you can go in an change the timestamp for logrotate.  The status of
logrotate is kept in /var/lib/logrotate.status.  It is a flat-file that
lists the logfile and the date it was last looked at by logrotate.

HTH
Jeff
-- 
Jeff Lovell
Sun Microsystems Inc.