Paul Jochum wrote: > Hi All: > > I am running the following: > > dom0: SunOS lss-nfsb01 5.11 snv_94 i86pc i386 i86xpv > with the following in my /etc/inet/ntp.conf: > [i]# --- OUR TIMESERVERS ----- > server lss-time1.ih.lucent.com > server lss-time2.ih.lucent.com > server lss-time3.ih.lucent.com > [/i] > > and the following command to start up the NTP service. > svcadm enable ntp > > domU: CentOS 5.2 image, with no updates yet. > > With ntp running on dom0, the time is accurate (within 200 milliseconds > normally, which is very acceptable). Both dom0 and all of the domU's are set > up for the same time zone, which is 5 hours behind UTC (Central Daylight Time) > > However, on the domU's (which are not running ntp), the clock are > consistently 5 hour behind the dom0 (again, within 200 milliseconds). To me, > it looks like Dom0 is not passing UTC to domU, but rather is passing the wall > clock to domU. > > So, is anyone else seeing this behavior, or did I make a mistake in my setup > somewhere? > > thanks, > > Paul
Solaris programs the "hardware clock" to be kept in UTC(GMT). When running on the hypervisor the hypervisor provides the simulated clock to the domU's. However, setting the time in your Solaris dom0 should set the time that the hypervisor provides to the domU's. I don't know how Linux expects the time to be kept in the hardware nor how often Linux checks the "Hardware" time after initially reading it. So you should at least be sure that your Dom0 has set the time correctly before starting your domU's. What is the output of the rtc command on dom0? (or alternately, what is the contents of /etc/rtc_config on your dom0? > > > This message posted from opensolaris.org > _______________________________________________ > xen-discuss mailing list > [email protected] _______________________________________________ xen-discuss mailing list [email protected]
