We have just purchased a cluster of 30 Sun x2100 servers in May to do
bioinformatics calculations. We are running CentOS 4.3 on all of them.
All x2100s have identical installations ntp syncing to the node
fileserver, a v240 running Sol10.
Here's the peculiar part:
7 of the x2100s have a clock drift of approximately 10 hours in the
first day. All 7 have a similar drift (+/- 200 seconds). Sometimes it
continues to drift more after it's about 10 hours ahead, but mostly it
stays at approximately 10 hours.
We have tried syncing them to a hardware clock once per day, we've tried
using different ntp servers. This happens with or without ntp enabled.
The motherboard has been replaced on one of the faulty 7 before we saw
this time problem, so at least one of the motherboards is new as of the
middle of June.
Does anyone have any idea what this could be?
I will summarize to the list.
Thanks for your help!
Matt Boes
UNIX Systems Manager
Max Planck Institute for Molecular Cell Biology and Genetics
_______________________________________________
sunmanagers mailing list
sunmanagers@xxxxxxxxxxxxxxx http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Re: Setting Up NTP Subnet ... > I have questions regarding best practices on architecture of NTP... > it sufficient to use multiple GPS receivers with ACTS dial ... The lower the stratum the bigger the ...servers and that the servers at this level should peer. ... (comp.protocols.time.ntp)
Re: Setting Up NTP Subnet ... > I have questions regarding best practices on architecture of NTP... > it sufficient to use multiple GPS receivers with ACTS dial backup ... me that we should create a stratum 2 tier peer layer a stratum 3 tier ...servers and that the servers at this level should peer. ... (comp.protocols.time.ntp)
Re: Hardware SNTP server ... But I certainly see what appears to be a pretty big drawback in the implementation, in that you're throwing away most of what I consider to be the best parts of NTP, just so that you can burn these things into FPGAs. ... Because these servers would not be the same and would not have precisely the same concept of time, to do this would be to completely destroy the NTP protocol itself. ... Even if they're all served by the same atomic clock reference, using serial cable splitters or whatever, those differences in cable lengths would be significant. ... What you want is broadcast, multicast, and manycast servers and clients, so that the load is automatically distributed across the set of servers, and the clients are automatically served by those machines which are close by. ... (comp.protocols.time.ntp)
Re: The libntp resumee... ... you have too much faith in ntp.... IF he is using his own servers (not outside ... using seven DIFFERENT poll intervals, one for each server because seven ... Currently we observe that both entry hosts can both become restricted due to ... (comp.protocols.time.ntp)
RE: NTP recommendations ... I then setup peer ntp on the chokepoint router to enable it to serve ntp to anyone requesting time. ... I only have the one hole in the firewall for the three destinations and DC active directory serves all clients, while the DMZ Router serves the rest. ... I am currently looking into configuring my company's time servers.... (Security-Basics)
