Understanding Node List and Heat Map Views
Applies To: Windows HPC Server 2008
In HPC Cluster Manager, you can view cluster nodes in two ways: the List view displays node properties and performance metrics in rows and columns, and the Heat Map view displays node performance metrics graphically. The two views provide customizable, at-a-glance information to help you monitor your cluster.
The node views provide a starting point for monitoring nodes, drilling down into node details, and managing nodes.
For information about using HPC Cluster Manager, see Overview of HPC Cluster Manager.
Node List view
In the List view, each row represents a node, and the columns display node properties and metric values. You can click the column headers to sort the list by specific properties. Customize the List view by choosing which columns to display: in the toolbar, click View, then click Column Chooser.
In the List view, there are three icons that may appear next to your nodes to indicate node health state:
A red cross indicates that the node is unreachable or that provisioning for that node failed.
A yellow warning sign indicates failed diagnostic tests for that node. For more information, see Resolve and Clear Diagnostic Alerts.
A blue circle indicates ongoing operations.
When you click a node in the List view, detailed information about that node appears in the Detail Pane. The Detail Pane displays information about the node such as hardware, operating system properties, and current performance metrics. Double-click a node to see this information in a separate window.
Node Heat Map view
In the Heat Map view, every cell represents a cluster node, and values for the tracked metrics are represented by a color gradient. You can choose the performance metrics to track from the Metric drop-down list. The Heat Map displays up to three metrics at a time.
The minimum value for a metric is associated with a color, for example, white, and the maximum value for that metric is associated with another color, for example, blue. In this case, lower values for that metric appear as lighter shades of blue, and higher values appear as darker shades of blue. For each metric, you can customize the maximum and minimum values and associated colors: in the toolbar, click Options, then click Customize View.
Double-click a node to see detailed information about the node such as hardware, operating system properties, and current performance metrics.
The Heat Map view displays the current performance metrics of a node. To view a graph of any of the performance metrics of a node over time, see View Performance Charts.
To add custom performance counters for your cluster, you can create an XML file that describes the metric to add, then import the new metric using the HPC PowerShell cmdlet
Import-HpcMetric. You can also modify an existing metric (for example, you can change the sampling frequency) by overwriting the existing settings. For more information, see Customize Metrics Collection in Windows HPC Server 2008.
To open HPC Cluster Manager, click Start, point to All Programs, click Microsoft HPC Pack, and then click HPC Cluster Manager. If the User Account Control dialog box appears, confirm that the action it displays is what you want, and then click Continue.