I recently built a Nixie tube clock with a kit purchased from http://www.pvelectronics.co.uk/ . They also have an "WiFi NTP Sync Device" which can be configured on WiFi to get time from an ntp server to keep the clock in sync. ("141 Units in Stock")
After setting it up, I immediately noticed it was polling the ntp server EVERY 6 SECONDS. This, to keep a clock with a 1 second resolution in sync. But, not only does the WiFi NTP Sync Device normally poll every 6 seconds, but it also does a DNS lookup each time it polls. While it's good to do a fresh DNS lookup on occasion, it's bad to do so for every poll. The canonical implementation of ntp, has a default setting for restricting incoming polls to an average of 8 seconds (https://www.eecis.udel.edu/~mills/ntp/html/accopt.html#discard). Many ntp servers send a "Kiss-o'-Death" packet when they find a client polling too often (https://www.eecis.udel.edu/~mills/ntp/html/rate.html#kiss). The PV electronics NTP thing does exactly the opposite of what it should do - instead of slowing down, it immediately does another DNS lookup and then does another NTP poll. So when it should be slowing down from its abusive 6 second polls, it actually speeds up to several polls per second, with a DNS lookup preceding each. Ouch! 6 second polls are, of course, abusive. They're also in direct violation of the ToS for the ntp pool servers, which PV Electronics suggests using. They were informed of these issues over 3 weeks ago (Pete Virica <p...@pvelectronics.co.uk>), and nothing has been done to address the issues. I don't know if anyone on the list has contact with Pete, or if anyone is in a position to push them to fix their stuff, but at this point I don't see any way to get them to react other than a good old public shaming. --- Mike _______________________________________________ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions