On Thu, Mar 03, 2016 at 10:54:47AM -0800, Grant Ridder wrote: > I noticed that one of my servers was several minutes fast. > > $ sudo chronyc -n sources > 210 Number of sources = 4 > MS Name/IP address Stratum Poll Reach LastRx Last sample > =============================================================================== > ^? 108.166.189.70 1 10 0 145d +6609us[+8115us] +/- > 10.6s > ^? 199.7.177.206 2 10 0 167d -25.7s[ +794us] +/- > 103ms > ^? 167.88.119.29 2 10 0 393d -343.5s[-4165us] +/- > 42ms > ^? 66.186.31.14 2 10 0 438d -410.8s[+6529us] +/- > 67ms
It looks like all four servers were shut down and removed from the pool. That's some bad luck. In the LastRx column you can see how many days ago the last response was received. Your server was drifting with no NTP synchronization for at least 145 days. The 10.6s value for 108.166.189.70 indicates the server was unsynchronized for some time before it stopped responding. A workaround is to restart the chronyd service, so it gets new IP addresses from the pool. The fix is to update chrony to 2.1.1 (available in CentOS 7.2), which can replace unreachable servers automatically. > $ grep ^server /etc/chrony.conf | cut -d" " -f2 > 0.centos.pool.ntp.org > 1.centos.pool.ntp.org > 2.centos.pool.ntp.org > 3.centos.pool.ntp.org -- Miroslav Lichvar _______________________________________________ pool mailing list [email protected] http://lists.ntp.org/listinfo/pool
