ntpd not functioning

[Waldher, Travis R]

> -----Original Message-----
> From: Rick Stevens [mailto:rstevens at internap.com]
> Sent: Wednesday, January 16, 2008 3:32 PM
> To: Getting started with Red Hat Linux
> Subject: Re: ntpd not functioning
> 
> On Wed, 2008-01-16 at 13:51 -0800, Waldher, Travis R wrote:
> > The error is:
> >
> >
> >
> > Ntpdate: no servers can be used, exiting
> >
> >
> >
> > My ntp.conf file has a “server 192.168.158.52” line in it.
> >
> >
> >
> > If I run “ntpdate -d 192.168.158.52” I get a valid response from the
> > ntp server so I know they are capable of talking to eachother.  Ntpq
> > even works.  But with or without the ntpd daemon running I am getting
> > the afore mentioned ntpdate error.
> >
> >
> >
> > I would like to just forget using ntpd since it’s usless in vmware,
> > but I can’t seem to get the ntpdate error messages to go away.
> 
> I can think of several ideas.  Perhaps port 123 is blocked by an
> iptables before ntpd is started and that restriction is later removed?

Sorry, I thought I stated it.  No iptables or hardware firewall anywhere.  As I said, when I run ntpdate with the debug switch I can talk to the time server.  If I take the debug switch off, I can't.  That is confusing me.

> 
> If you don't want ntpd at all, why not just "service --level 2345 ntpd
> off" in the vmwared host?  ntpdate is run by the ntpd startup script.

I would love to, but, VMWare does a HORRIBLE time of keeping time on a system and for some reason ntpd is unable to keep the time sync'd at all.  It may have to do with the virtual CPU issues causing the time sync problem in the first place.

On one set of systems, to keep the time within even 2-3 seconds of eachother I have a cron job executing an ntpdate every 15 seconds.  That's how fast it races.

> 
> Did you also check /etc/ntp/step-tickers and /etc/ntp/servers and make
> sure they're empty or commented out?

Not empty or commented out.  Could feeding it the IP address and not a host name be causing problems?