This is a relatively simple configuration. All we need to know are the addresses of the timeservers we have on our network; the timeservers can be any type of server, even a local clock impersonator. In this case, our timeservers are hpeos002 (reverted back to be a Stratum-2 server) and hpeos003 . I just need to set up the files /etc/ntp.conf and /etc/rc.config.d.netdaemons . As before, the changes are not difficult in this situation: root@hpeos001[] # vi /etc/ntp.conf ... server hpeos002 server hpeos003 driftfile /etc/ntp.drift root@hpeos001[] # vi /etc/rc.config.d/netdaemons ... export NTPDATE_SERVER="hpeos002 hpeos003" export XNTPD=1 export XNTPD_ARGS= We can now start the daemons as before: root@hpeos001[] # /sbin/init.d/xntpd start 28 Aug 11:15:32 ntpdate[19404]: step time server 192.168.0.202 offset 1238071.026636 sec xntpd root@hpeos001[] # Immediately, we should see a step-change in the time of this machine. After the usual 5 minutes while the daemon calibrates itself, we should see a settling down of the daemon, keeping us synchronized with our timeservers. root@hpeos001[] # date ; remsh hpeos002 date Thu Aug 28 11:17:36 BST 2003 Thu Aug 28 11:17:36 BST 2003 root@hpeos001[] # root@hpeos001[] # ntpq -p remote refid st t when poll reach delay offset disp ============================================================================= +hpeos002 192.6.38.127 2 u 5 64 377 1.19 -4.147 0.63 *hpeos003 192.36.133.17 2 u 21 64 377 0.64 -4.102 0.34 root@hpeos001[] # |