Suggestion on Parent / Child hosts in relation to Services

[JakeHatMacys]

So here's a cliff note on what I'm looking at...

I have a Switch which is a parent and a device which is the child. I then have an application port check which is a "Service" check on the child.

From what I've seen the Parent / Child host relationship does not transfer to the services. So if I have 30 children on the Switch I'm going to ticket on 30 services being down if the switch goes because the switch is just taking down "host" alerting on the children. Do I have this correct?

If so is there an easy way to apply this relationship to services? And oh by the way I need to do this 800 times! *UGH* The plan is to use the Bulk copy and originally was looking at service dependencies in Core but they're not copying when I do the buik move.

Thoughts suggestions? Wondering if there's some Group rules I could use.

[lmiltchev]

So if I have 30 children on the Switch I'm going to ticket on 30 services being down if the switch goes because the switch is just taking down "host" alerting on the children. Do I have this correct?

When the host goes to "hard" down or unreachable state, you shouldn't be getting notified about services on this host.

You can choose not to receive the notifications on "unreachable" states, in order to "reduce the noise".

http://nagios.sourceforge.net/docs/nagi ... ility.html

[JakeHatMacys]

I think the problem is the Host isn't incrementing to a hard state... And I have no idea why:

I have the check set for every 2 minutes and then 1 minute after an error FYI. Are their any known bugs with this?

[slansing]

No, there are no known bugs with this, can you post your configuration file for this host, as well as it's template if it has one assigned? CCM > Hosts > Click the diskette icon next to the host's name. What version of XI is this?

[JakeHatMacys]

No, there are no known bugs with this, can you post your configuration file for this host, as well as it's template if it has one assigned? CCM > Hosts > Click the diskette icon next to the host's name. What version of XI is this?

define host {
host_name cvt3247532.federated.fds
use xiwizard_genericnetdevice_host
address 11.48.xxx.xxx
parents 11.51.xxx.xxx (removed because I'm paranoid)
hostgroups CVT Test
max_check_attempts 5
check_interval 2
retry_interval 1
active_checks_enabled 1
check_period xi_timeperiod_24x7
notification_interval 60
notification_period xi_timeperiod_24x7
notification_options d,u,
icon_image network_node.png
statusmap_image network_node.png
_xiwizard genericnetdevice
register 1
}

And the parent is in a green state.

[JakeHatMacys]

More on what I'm seeing, the problem will start after it hit's 5/5.... it does on the first go around sends a snmp trap but then resets after about 2 minutes to 1/5 and doesn't move. (The host that is) So now the service alerting won't stop and we'll alert on the service after it tries 50 times in one minute intervals.

Is there a setting that would cause this? I'm testing on a CVT in my cube and literally unplugging it. So there's not way it could go green for a blip.

[lmiltchev]

it does on the first go around sends a snmp trap but then...

Can you elaborate on this? The config you posted seems to be for an active check...

Code: Select all

max_check_attempts 5
check_interval 2
retry_interval 1
active_checks_enabled 1

[JakeHatMacys]

I had actually went into core and swapped that on after it wasn't incrementing. I was really just trying things at this point.

Going to delete the monitor & host recreate it and see if anything changes...

[JakeHatMacys]

Recreated it, seems to be the same issue:

define host {
host_name cvt3247532
use xiwizard_genericnetdevice_host
address cvt3247532
max_check_attempts 5
check_interval 2
retry_interval 1
check_period xi_timeperiod_24x7
notification_interval 60
notification_period xi_timeperiod_24x7
icon_image server2.png
statusmap_image server2.png
_xiwizard tcpudpport
register 1
}

And here is the service running on the host:

define service {
host_name cvt3247532
service_description TCP Port 25000
use xiwizard_tcp_service
check_command check_xi_service_tcp!-p 25000
max_check_attempts 50
check_interval 2
retry_interval 1
check_period xi_timeperiod_24x7
notification_interval 60
notification_period xi_timeperiod_24x7
notifications_enabled 1
contacts null
contact_groups null
_xiwizard tcpudpport
register 1
}

[JakeHatMacys]

Also the version of XI I'm running is:

Nagios XI 2014R1.4