Disable ping check and only check by open port is host is up

  • Do need an extra rule for hosts that are reachable over a dyndns name?

    Actually not, if you defined the dns name as ip address in the host properties.

    We need to check if the core uses the right dns name as address while doing the check.

    The check command uses the $HOSTADDRESS$ macro which is filled by the address directive in the nagios config files (if you are using nagios as core for your OMD site). To verify the correct address, run the following command:


    Code
    1. # cat /omd/sites/yoursite/var/nagios/objects.cache | grep address.*yourdnsname

    The output should show you


    Code
    1. address yourdnsname
  • Yes, the core ist using that name:

    Code
    1. root@checkmk:~# cat /omd/sites/mysite/var/nagios/objects.cache | grep addressofhost.*dyndnsprovider.de
    2. members checkmk,esx1.mysite.de,esx2.mysite.de,esx4.mysite.de,s01.schule.local,zcm01.schule.local,ebehumbolt.dyndnsprovider.de,addressofhost.dyndnsprovider.de
    3. host_name addressofhost.dyndnsprovider.de
    4. host_name addressofhost.dyndnsprovider.de
    5. check_command check_mk_active-tcp! -p 6557 -H 'addressofhost.dyndnsprovider.de'
  • Okay, very strange. I delete the host and recreated it. Now one of them is working probably. The state is up now and only TCP Port 6557 is being checked:


    The other host (same configuration in WATO - just another dyndnsname) with the same TCP-Port check is marked as down. - Also after recreating it:


  • The other host (same configuration in WATO - just another dyndnsname) with the same TCP-Port check is marked as down. - Also after recreating it:

    But the service is running correctly as your second picture shows.


    Can you confirm that your dns server is resolving the dyndns host names are consistently?

  • Okay, after I entered the status host under distributed Monitoring to the site it work and the state of the host move to up!

    The strange thing is that other hosts, where I also haven't set the status host under distributed monitoring were in the state up before.

    Anyway this worked for me :)