> I don't know why it didn't manage to get its clock synced, despite > sending NTP requests every 32 seconds. It may well have lost sync exactly _because_ it was sending NTP requests every 32 seconds.
Many NTP servers will death-kiss and then refuse to service clients that keep sending NTP requests more often than reasonable. I understand that client was originally requesting time service too often from your server. How is your server configured in this respect? When the sever stops servicing them, the clients often follow the route "if it doesn't work, try more of the same". Which (of course) doesn't work. Such clients keep up the bad work of sending packets too fast, so in consequence they never get un-ignored. When you reset the NTP configuration, the client probably switched to a different server. Which may at some point in the future also send a kiss of death and start to ignore it. In the meantime, the first few packets which that other server did service allowed the client to sync. _______________________________________________ pool mailing list [email protected] http://lists.ntp.org/listinfo/pool
