[OpenAFS] Howto prevent afs from resetting the system clock?

Martin Schulz schulz@iwrmm.math.uni-karlsruhe.de
03 May 2001 10:26:39 +0200


"Neulinger, Nathan" <nneul@umr.edu> writes:

> > It is a bad idea to run ntp on AFS clients. The cache manager 
> > synchronizes
> > the clock with the AFS servers, which should run runntp. I 
> > use the following
> > configuration:
> 
> No... that's not true... real ntpd does a LOT better job than afsd does.
> What is a bad idea is running both. Ideally, you should probably be running
> NTPd on both your afs fileservers and your clients.

That is what I am doing (or at least I think that I do so). 

I have a bad feeling about using afs for solving every problem under
the sun. That approach of "one-stop-installation" is ok from a
commercial point of view, where it is helpfull for the support people
that all components come from one controllable source. 

However, I wish my afs installation to concentrate on its core
business, that is file serving, preparing backups etc. I already had
the (x)ntpd daemon running before so why should I switch over to a
nonstandard one?

Thanks to all for the massive-parallel answers to my question. 

I would love to see an (more or less) bare-bone openafs distribution
(e.g. without kas and runntp but with native support for krb5) that
installs easily into an environment of existing framework of other
services. One can always dream of something...

Yours,
-- 
Martin Schulz                             schulz@iwrmm.math.uni-karlsruhe.de
Uni Karlsruhe, Institut f. wissenschaftliches Rechnen u. math. Modellbildung
Engesser Str. 6, 76128 Karlsruhe