trey at ERC.MsState.Edu
Thu Oct 11 14:35:19 EDT 2001
>From: Greg Lindahl <lindahl at conservativecomputer.com>
>On Thu, Oct 11, 2001 at 07:07:34PM +0200, Jakob Østergaard wrote:
>> It was the "definitive" statement that I objected to.
>Given that the question was about "a fairly large cluster with a
>fairly small number of users", I think my answer was correct about a
>majority of the installations in the field. You didn't notice anyone
>with a large cluster leaping up to disagree, did you? That's how we
>take polls in our community ;) ;)
Ok, so I guess I should leap. We have a 292 node (584 processor)
cluster that utilizes NIS, and for the most part it works well. We
wanted our cluster to integrate with our existing systems (hundreds of
Sun and SGI machines) as seemlessly as possible and NIS was the
solution that we chose.
As with everything else, I think that the choice to use NIS is very
much site and application dependent. In some cases NIS makes more
sense than straight files.
Trey Breckenridge - Computing Systems Manager - trey at ERC.MsState.Edu
Mississippi State University Engineering Research Center
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
More information about the Beowulf