Re: [cobalt-users] Raq3 vs. real-time clock

[Mailing Lists <listonly@xxxxxxxxxxxxxxxxxxxx>]

on 7-5-02 12:50 PM, Jolley, Carl stated:

> Due to tales I've heard about the problems with the realtime
> clock, I've put a call to ntpdate near the end of
> /etc/rc.d/rc.local but by the time it happens, several cron
> jobs have already started. Can I put a call to ntpdate somewhere
> in /etc/rc.d/rc.sysinit? I'm not getting mail messages saying
> the clock is messed up but obvisously the clock is getting
> set wrong at boot time. In addition to tweeking my startup
> scripts, should I ask my co-loc vendor to replace the battery?
> Will this do any good?

We did the battery thing, didn't help. You have an interesting point on the
crons running. I had brought that up to a cobalt engineer when they were in
my box trying to correct the problem. He stated that cron was to smart. I
told them that once the cron has ran for july10 even though the date was
only july5th, that cron would not run again until that date had passed. We
argued for a while on that. He did state that the Raq3's had some type of
ROM issue on the clock hardware. His other statement was I had a motherboard
puking. I have yet to resolve the issue. I may try Gerald's idea from today
on NTP server.

Dave