I have a strange issue... I have a host in DMZ which need to be monitored using NRPE. NRPE daemon starts fine with out any problem.. When i fire a check command using check_nrpe from the nagios server, it spawns a new nrpe process.but in the nagios server check timesout after 10 seconds. but more strangly the new spwaned process doesnot die by itself and remains there.So for each check there is a new process created and it does not terminate by itself. So after few hours you can see thousands of nrpe processes and it kills the box.
nrpe works fine for all the other hosts which are not in DMZ. So looks like tough check_nrpe can connect to port 5666 on the remote host ,somehowe its unable to sent the check results back.That may be reson why the check_nrpe timesout after 10 second. I am not sure how exactly nrpe works.
Another observation i have seen is traceroute from remote host to nagios server fails since this is in a DMZ. Does nrpe do any kind of icmp check when it send check results back to check_nrpe ? Why is the newly spawned process does not kill itself?
Any help on this is appreciated.
NRPE version is 2.12