Cluster Monitoring

[mjandl]

Hello,

we are running a Veritas - cluster consisting of 2 OracleLinux nodes.
How would we need to configure which rules so that Veritas cluster resources are correctly represented.

E.g. VCS Resource DiskGroup_01 on cluster cluster: db-cl.
detailed output of the plugin:
[HOST01]: offline WARN
[HOST02]: online

should show online in my perception

[HOST01]: offline WARN
[HOST02]: faulted

should show e.g.: CRIT.

Has anyone already implemented this in their CHECKMK environment and could give me a little guidance on this?

CHECKMK Version: Enterprise 2.0.0p8

Greetings,
Mario

[mike1098 (Michael FRANK)]

Hi Mario,

First you need to create a rule “Clustered services” and add all service which are clustered like your Diskgroups.
Then you create in Hosts a new object with "New Cluster " and add your cluster nodes to this object.
In the services you will see then the prior configured services. This services are OK as long as ONE Service on the nodes is OK.
Also, this services vanish from the host objects of the nodes. You will see a list of the clustered services in the WATO Service Discovery view at the bottom.

Hope that helps

Michael

[mjandl]

This behavior I have expected. But this did not happen. I have recreated this scenario in a new environment.

A few more details:

Host 01 & Host 02: Oracle Linux with CHECKMK Linux Agent installed.
Veritas Cluster software installed on both hosts.

Agent output for Veritas diskgroup IMPAXIMPFSSG_DiskGroup_impfsdg:
Host 01: IMPAXIMPFSSG_DiskGroup_impfsdg State dbs1-pacs OFFLINE.
Host 02: IMPAXIMPFSSG_DiskGroup_impfsdg State dbs2-pacs ONLINE

Rule clustered services for Host01 and Host02 created and all Veritas resources allocated.
Cluster host CL01 created.

Output for IMPAXIMPFSSG_DiskGroup_impfsdg on CL01:

In my opinion, the cluster implementation is designed for the worst case.
However, this is not purposeful here.

Did I miss something in the configuration?

Greetings,
Mario