hosts shown down when actually up

[Jeff Kilpatrick]

I have setup Nagios to monitor around 135 hosts (only checking to see if host is up).  My network topology is a MAN using Gig Ethernet from the site where the Nagios machine is and a mix of 10/100/1000Mb/s ethernet to each site on the MAN.  I am monitoring the router connecting each site as well as 3-4 servers at each site.  One site that is connected via 100Mb/s will show the router down and/or some of the servers down; sometimes all, sometimes different ones.  Yet, when one of these machines is shown as down, I am able to ping it from the nagios machine and other machines.  I can restart nagios and it will show these machines as up for a while, then return to a down state.  I'm also monitoring the performance of these using MRTG which currently is reporting one machine at this site as being up for 16 days while Nagios reports it has been down for 3 days.  I have included the service definition below; the hosts definitions are identical to the other hosts which are working fine.  The router at this site is in a hostgroup with all the other routers; the servers are in a hostgroup unique to this site.  
  Also should note that this particular site did actually go down once after I restarted nagios; nagios reported it was up, when the site was down.  
  Any help in troubleshooting this is greatly appreciated.  
  

define service{
        use     generic-service ; Name of service template to use

        host_name                       *
        service_description     PING
        is_volatile                     0
        check_period            24x7
        max_check_attempts      3
        normal_check_interval   5
        retry_check_interval    1
        contact_groups          router-admins
        notification_interval   120
        notification_period     24x7
        notification_options    c,r
        check_command           check_ping!100.0,20%!500.0,60%
        }
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.monitoring-lists.org/archive/users/attachments/20030602/ee5599ac/attachment.html>