Looking at this page: http://www.gentoo-portage.com/net-misc/ntp/ChangeLog#ptabs 09 Jan 2007; Mike Frysinger <vapier@gentoo.org> files/ntp-client.confd, files/ntp-client.rc, ntp-4.2.2_p3.ebuild, ntp-4.2.2_p4.ebuild, ntp-4.2.4.ebuild: Cleanup ntp-client script by scrubbing the lame timeout code that never really worked. NTP itself handles timeouts sanely now. The problem is that NTP doesn't handle DNS timeouts very well due to network problems. With two ntp servers configured and the network connection broken, system startup will be delayed by a full 2 minutes. I'm not sure what Mike meant by 'never really worked' seeing as I've replaced the code in my local init script and tested it and it seems to function as designed. Giving it a NTPCLIENT_TIMEOUT of 10 seconds, breaking the network connection and restarting the machine results in ntpdate being killed after 10 seconds as expected. The built in ntpdate timeout settings do not seem to apply to DNS lookups, only the actual NTP query itself. Reproducible: Always Steps to Reproduce: 1. Add 2 servers to your /etc/conf.d/ntp-client 2. Break your network 3. Reboot and watch with joy as your startup process now takes an extra 2 minutes The original functioning init code: ebegin "Setting clock via the NTP client '${NTPCLIENT_CMD}'" ${NTPCLIENT_CMD} ${NTPCLIENT_OPTS} >/dev/null & local pid=$! (sleep ${NTPCLIENT_TIMEOUT:-30}; kill -9 ${pid} >&/dev/null) & wait ${pid} eend $? "Failed to set clock"
configure dns timeouts in your resolv.conf. see the man page for more information.
