If you are peering with a machine that goes down, or that stops running an NTP daemon, you will see screams in your syslog of the form: xntpd[26005]: recvfrom() fd=7: Connection refused Worse, these are logged with priority LOG_ERR, which is supposed to mark fairly important problems. These exist because Linux + glibc return errors on recvfrom() that are due to a previous UDP sendto() on the same socket failing because of various detectable problems. The xntp3 code is not expecting this to happen, and thinks that these are real recvfrom() errors. These syslog messages are annoying. I've developed a patch to muzzle them for connection refused, host down, and host/net unreachables, if people are interested.
Could you reply to this message with patch attached? Thanks ...
closed due to lack of additional feedback, and the fact that this is solved with newer versions in currently available distributions of Red Hat Linux.