On Tuesday 05 March 2019 at 17:59:47, Ivo Kidze wrote:
> Hi Anthony,
>
> Can you send a sample configuration?
My Icinga2 configuration for these machines is pretty complex, because I use a
lot of templates and apply rules to define things once and use them multiple
times.
However, the key to it all is that I have a Host definition such as:
object Host "Dev Cluster" {
import "ClusterDefs"
vars.cluster_nodes = [ "node1.cluster.tld", "node2.cluster.tld" ]
}
Those node1.cluster.tld and node2 are real machines with an IP address, and
running services which get checked by Icinga.
The template ClusterDefs contains a dummy check command which looks at the
host states of the two nodes in the cluster, to work out what state to report
for the cluster itself.
Similarly I have a service template for cluster machines, which looks at the
last service state of each node, and reports an appropriate result for the
cluster as a whole (depending on what the check is, I might want the maximum
value, the minimum value, the sum of the values, etc).
So, what I want to be able to do is to put the "Dev Cluster" machine into
Downtime, and have this take effect on node1.cluster.tld and node2.cluster.tld
as well, so that I get no notifications from any of them until I take the "Dev
Cluster" machine out of Downtime again.
I'm wondering if I can manage this with Icinga Dependencies, but somehow I
don't think so, because that seems to be more for defining "if this host is
down, don't tell me about those other two hosts" (for network connectivity
reasons etc), whereas I *do* want to get all the notifications which apply when
problems occur, *except* when I've explicitly put something into Downtime.
Does that help explain the scenario?
Antony.
> -----Original Message-----
> From: icinga-users [mailto:[email protected]] On Behalf
> Of Antony Stone Sent: Friday, March 1, 2019 12:21 PM
> To: Icinga User's Corner <[email protected]>
> Subject: [icinga-users] Single-click downtime for multiple hosts (and all
> associated services)
>
> Hi.
>
> Background:
>
> I have a high-availability setup which consists of three machines running
> corosync / pacemaker managing a floating IP address, and a second group of
> three machines similarly running another floating IP address.
>
> Either of those addresses can be used to access various services running on
> the entire "cluster".
>
> In Icinga(2.10.3) I've created host entries for:
>
> - each of the six "real" servers
> - each of the two floating IPs at which services can be accessed
> - one "cluster" entity which represents either of the floating addresses
> being available
>
> Requirement:
> From time to time I run updates to the software on one or more of the real
>
> servers, and of course this results in short-term unavailability of the
> services they are running, during the upgrade.
>
> If I'm updating just one machine, I can put it into Downtime in Icingaweb2
> (along with all its services) and avoid getting alerts while the services
> (or indeed the host, if a reboot is involved) are unavailable.
>
> I would like to be able to set Downtime on one of the floating IP "Hosts"
> and have it take effect on all three of the real servers which provide
> this single address, or to set Downtime on the single "cluster" Host, and
> have this take effect on all six servers together.
>
>
> Can anyone think of a way to do this?
>
> I've tried using Group membership, and custom Host variables, but I can't
> find any way to have a setting I change on the "higher level" host ripple
> down to the component hosts which are members of it.
>
>
> Any ideas gratefully received :)
>
>
> Antony.
--
Pavlov is in the pub enjoying a pint.
The barman rings for last orders, and Pavlov jumps up exclaiming "Damn! I
forgot to feed the dog!"
Please reply to the list;
please *don't* CC me.
_______________________________________________
icinga-users mailing list
[email protected]
https://lists.icinga.org/mailman/listinfo/icinga-users
