checks timeout when DNS has problems

[lyle]

When our DNS servers have problems, we get timeouts on a lot of checks (I think mostly on Windows hosts using one of the clients via NRPE). We then get notifications up the ying-yang, and more notifications when DNS gets fixed.

At first glance, it would seem that hosts are all defined on the Nagios server along with their IP addresses, so this wouldn't happen.

I'm guessing the request gets sent out, but responses can't get back to the server because the client needs to do a lookup on the server's name. Odd, though, because "allowed hosts" in the client config is usually specified as an IP address, not a hostname. Maybe the Server's name, not it's IP address, is embedded in the request.

Any insight? Thanks...Lyle

[mguthrie]

Just to clarify, are the remote hosts being referenced by IP address, or DNS?

Does it make any different to increase the timeout ( -t ) value in the check commands?