Monitor cluster with tng 2.4 agent

[amiga512]

Hi,

We have some clusters using agent 2.4 from TNG.
We need to monitor some processes on each node of the cluster.
In one node process are up and in the other node process are down.

We have agents on the both nodes.

How whe could monitor only the virtual node who is associate to the process, or to freeze "process down" messages when the cluster if failover the other node ?

Thanks You

 

[unitek]

Hi

Our solution to this problem is to have different configsets that are automatically loaded when a given trap is recieved. eg The clustering software fails over from server A to server B, have the clustering software send a "failover" trap to your console indicating this, pick up the trap and load the configset with minimal components on server A and then load the configset with all components on server B.

Prior to that we were using adminstat to unmanage whole agent objects on the passive server of the cluster pair. The first solution is the preferred one.

Hope this helps!

 

[gmorg]

I have also set up config sets to deal with this, what I have done is have one config set, defining
"AWS$AGENT_VMS_CONFIGSET" = "node1_LOCAL"
"AWS$AGENT_VMS_CONFIGSET_CLUSTER" = "node1_CLUSTER"

and on the other system defining node2_local & node1_cluster.

this way you only need to create one config set, which the
appropriate details for each system, the details for the
cluster go into the cluster definition.

 

[amiga512]

Thanks You

We decided to use to use CMO product from CA
cause we have 200 Clusters now whith differents configs.