Not all hosts working on satellite

Hi,

right now I’m experiencing some odd behaviour from a single icinga2-satellite:

When running icinga daemon -C everything seems fine and the satellite confirms 42 hosts, but only 24 of them are shown on icingaweb2. Everything worked fine until we moved the hosts to the (for them) local satellite.

Affected are the hostconfigs in conf.d/<subfolders>.
include_recursive "conf.d" is active in icinga2.conf.

To clarify the obvious question: hosts which add themselves to icinga2 via puppet are located in zones.d and work fine. Only the manually configured hosts have issues right now.

Other zones and satellites which are setup similarly and are doing fine.

Last lines of icinga daemon -C:
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 2 NotificationCommands.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 1 IcingaApplication.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 69 HostGroups.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 42 Hosts.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 4 EventCommands.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 67 Downtimes.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 2 Comments.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 1 FileLogger.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 1 CheckerComponent.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 1 ApiListener.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 6 Zones.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 2 Endpoints.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 1 ApiUser.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 261 CheckCommands.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 169 UserGroups.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 62 Users.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 3 TimePeriods.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 38 ServiceGroups.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 42 ScheduledDowntimes.
[2022-02-04 11:44:33 +0100] information/ConfigItem: Instantiated 687 Services.
[2022-02-04 11:44:33 +0100] information/ScriptGlobal: Dumping variables to file ‘/var/cache/icinga2/icinga2.vars’
[2022-02-04 11:44:33 +0100] information/cli: Finished validating the configuration file(s).

Does anyone know why this may happen?
Kind regards and thank you in advance!
Markus

To me it looks like your host objects are not well configured at your master.

BTW: Your satellites know about 2 endpoint object only. Do you have icinga’s agent installed on the hosts? If so, you need endpoint and zone objects accordingly.

Hi Roland,

The satellite only knows 2 endpoints: itself and the master-node.
The zones do not know about other zones except the master-Zone.

It’s strange to me, that hosts in satellite-zones are never configured on master and it works everywhere except in this particular zone.

Kind regards,
Markus

zones.conf on satellite:

object Endpoint “satellite” {
host = “x.x.x.x”
port = 5665
log_duration = 1d
}

object Endpoint “masternode” {
host = “x.x.x.x”
}

object Zone “sat” {
endpoints = [ “satellite”, ]
parent = “master”
}

object Zone “master” {
endpoints = [ “masternode”, ]
}

object Zone “director-global” {
global = true
}

object Zone “global-templates” {
global = true
}

object Zone “linux-commands” {
global = true
}

object Zone “windows-commands” {
global = true
}

Ah, you’re using the director, than it’s a different story since it handles zone and endpoint objects for the hosts if needed.

If the master does not know about hosts e.g. defined at a satellite only, the’ll never appear in icingaweb. You can check your host objects with icinga2 object list -t host for all host or with -n <hostname> in addition for a particular host object. And both master and corresponding satellite needs to have a object with the same name.

1 Like

Thank you, that seems to be the solution!

It seems that the other zones are only working due to somhow bugged configurations, which seem to be still available on the master node (even thou they’re deleted).
After I added the host-configs to the master node everything is working fine!

Have a great weekend!
Kind regards,
Markus