HA installation - documentation clarification

Icinga 2
1

Hi, I need to install a new node in the master zone of my current setup. I have come across a quiestion about the installation document:

It is clear that I need to stop my current monitoring master and then reconfigure both nodes and replicate var-lib-api-involved-files, so that the new HA cluster in the master zone keeps current monitoring state (downtime, passive checks, host & service state).

However, following question arises:

Question 1. When starting up the cluster in the master zone, shall icinga2 service be started on both cluster members at the same time? or shall one of them to be started first and wait for log-alerting about its peer not connected before starting the other? (is it safe to assume that both nodes can be started simultaneously from the technical concepts document ?)

Question 2. When reconfiguring all other endpoints/Zones to accept connections from the new master, Do all Endpoints have to be stopped as on the Quoted document PRIOR to starting the master cluster?
Or should the reconfiguration of the child Endpoints be made AFTER the master zone is running in the HA cluster.

Quoting the document:

Note

Ensure that all endpoints are shut down during this procedure. Once you have synced the cached files, proceed with configuring the remaining endpoints to let them know about the new master/satellite node (zones.conf).

Output from icinga2 daemon --validate (80 Endpoints)

[2021-02-04 15:38:33 +0000] information/cli: Icinga application loader (version: 2.11.6-1)
[2021-02-04 15:38:33 +0000] information/cli: Loading configuration file(s).
[2021-02-04 15:38:34 +0000] information/ConfigItem: Committing config item(s).
...
1-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 110 ScheduledDowntimes.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 299 HostGroups.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 1 EventCommand.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 1 FileLogger.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 3 NotificationCommands.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 1 NotificationComponent.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 97545 Notifications.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 1 IcingaApplication.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 6246 Hosts.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 113432 Dependencies.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 1 CheckerComponent.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 66 Zones.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 80 Endpoints.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 1 ApiUser.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 425 CheckCommands.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 6 TimePeriods.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 2 Users.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 88340 Services.
[2021-02-04 15:40:09 +0000] information/ConfigItem: Instantiated 51 ServiceGroups.
[2021-02-04 15:40:09 +0000] information/ScriptGlobal: Dumping variables to file '/var/cache/icinga2/icinga2.vars'
[2021-02-04 15:40:09 +0000] information/cli: Finished validating the configuration file(s).

Thanks in advance,

2

Hi,

you don’t have to restart the nodes simultaneously. You can do one node after the others. If the config files are changed in the correct way, the nodes find each other.
Think about maintenance (Updates), restarts, connection losts because of network failure etc. :wink: In principle it’s the same.

3

Hi Sy,
Thanks for your prompt reply.

I’ll asume that child zones should be reconfigured and restarted prior to master zones (Q2) so both masters are able to connect (zones are configured one-way from master to child); and then, the master cluster can be restarted in any order.

Right?
Best,

4

Generally, yes. But I personaly would restart with the config master first. But you don’t need to restart the hole server. You only have to restart the icinga2 daemon.
Don’t forget to reconfigure the zones.conf also on the master1