Not content with watching clocks drift on servers, I thought it might be prudent to keep an eye on such things. We already monitor our NTP server drift with Nagios, so why not extend it to all of our Linux servers?
The only catch here is that not all servers run the NTP server, but they do run the client. In other words, there’s no NTP port in which to poke the server on. The next best solution then is to compare the time available over SNMP with the local time of the local server (which is kept accurate by NTP)
Thankfully someone got there before me… using the setup described at http://meinit.nl/nagios-time-check-using-snmp meant that comparing the local time with NTP time was done across 50 servers in approximately 15 minutes. Nagios saves the day yet again.
Now, time to deploy an updated ntp.conf to 20 odd servers!
Recent Comments