A.2 Planning to Monitor Cluster Resources

Before installing agents for cluster monitoring, you should prepare for the deployment. As with any AppManager-related planning, first evaluate the characteristics of your environment and your monitoring needs.

The questions you need to answer include:

  • What resources do you need to monitor in MSCS?

  • Which applications are installed as clustered applications?

  • Which Knowledge Scripts will you need to run to effectively monitor a given cluster as a virtual server?

  • How many physical nodes make up each MSCS?

  • What resources have you defined for each cluster group?

  • Are multiple instances of the application you want to monitor running on the same node?

  • Is the cluster configured as an active/passive cluster or as an active/active cluster? If active/active, how many drive resources can independently fail from node to node?

Create a worksheet for tracking the information you gather, particularly if you are monitoring more complex cluster configurations. For example, if MSCS relies on six or eight physical nodes:

  • The Cluster Name resource and Quorum Disk should be located in the same cluster group, thereby guaranteeing that they are active on the same node at any given time. Because the cluster stores resource information on the MSCS directory on the Quorum Disk, the cluster operation fails if you take the Quorum Disk offline.

  • If the cluster uses multiple instances of the application server and is configured as an active/active cluster, setting up the AppManager agent on each node and properly configuring the jobs to handle failover on each node requires careful planning.