NTP Time Sync on Proliant DL580

[Lincoln Deslauriers]

I have some Compaq Proliant servers which are running NTP and synchronizing to a Lantronix NTP server connected to a GPS clock. The
clock on the server remains synchronized to within a couple of ms for most of the time. However, occationally (at least once a day) the server clock begins to deviate wildly from the NTP source (see the attached file).

The same problems also happens the other way around, where DAC1 stays synched and DAC1 skews wildly.

I've tried to shutdown any other programs I could think of which might be adjusting the time (i.e. the Windows 2000 time service). Anyone have any suggestions as to what may be causing this?

Thanks,

Lincoln

[Claude Boileau]

Hi Lincoln

Here are three links I found if not already check out
I gather RTP is for (Network Time Protocol):

http://support.novell.com/cgi-bin/search/tidfinder.cgi?10011518

http://support.novell.com/cgi-bin/search/tidfinder.cgi?2956780

http://support.novell.com/cgi-bin/search/tidfinder.cgi?2926204

Hope this will help

Claude

[Lincoln Deslauriers]

The servers are running Windows 2000. NTP seems to be running fine most of the time but occasionally (maybe once a day or so) the offset climbs from 1 or 2ms to up to 1000ms for no apparent reason. Could the clock on the server really be this bad or might there be other forces at work here?

[Tele Charalambous]

I am experiencing the same kind of problems with DL380 G3's, where the actual display of the clock seems to be running fast for 5 seconds, then pausing, then speeding up, etc etc. Ive actually seen the clock deviate by days over an extended period! Although I cannot find anything in the forums or support, this also seems to affect programs I am trying to install (eg Oracle Client 8.1.7 installer doesnt even run! Weblogic 5.1 install causes a Dr Watson!!)

This has happened on 2 identical 380G3s with Dual Xeon 2.4GHz processors servers with Windows NT4, they both had 2.5GB RAM (which I have tried dropping to 500MB, with no difference).