Monitoring networks by using GUI

The Network page provides an easy way to monitor the performance, health status, and configuration aspects of all available networks and interfaces that are part of the networks.

A dedicated network is used within the cluster for certain operations. For example, the system uses the administration network when an administration command is issued. It is also used for sharing administration-related information. This network is used for node-to-node communication within the cluster. The daemon network is used for sharing file system or other resources data. Remote clusters also establish communication path through the daemon network. Similarly, the dedicated network types like CES network and external network can also be configured in the cluster.

The performance of network is monitored by monitoring the data transfer that is managed through the respective interfaces. The following types of network interfaces can be monitored through the GUI:
  • IP interfaces on Ethernet and InfiniBand adapters.
  • Remote Direct Memory Access (RDMA) interfaces on InfiniBand adapters with Open Fabrics Enterprise Distribution (OFED) drivers.
The GUI retrieves performance data from the performance monitoring tool. The IP-adapter-based metrics are taken from the Network sensor and the RDMA metrics are taken from the InfiniBand sensor. If no performance data appears in the GUI, verify that the monitoring tool is correctly set up and that these two sensors are enabled.

The Network page also exposes adapters and IPs that are not bound to a service, to provide a full view of the network activity on a node.

The details of the networks and their components can be obtained both in graphical as well as tabular formats. The Network page provides the following options to analyze the performance and status of networks and adapters:
  1. A quick view that gives graphical representation of overall IP throughput, overall RDMA throughput, IP interfaces by bytes sent and received, and RDMA interfaces by bytes sent and received. You can access this view by selecting the expand button that is placed next to the title of the page. You can close this view if not required.
    Graphs in the overview are refreshed regularly. The refresh intervals of the top three entities depend on the displayed timeframe as shown:
    • Every minute for the 5-minutes timeframe
    • Every 15 minutes for the 1-hour timeframe
    • Every 6 hours for the 24 hours timeframe
    • Every two days for the 7 days timeframe
    • Every seven days for the 30 days timeframe
    • Every four months for the 365 days timeframe

    If you click a block in the IP interfaces charts, the corresponding details are displayed in the IP interfaces table. The table is filtered by the IP interfaces that are part of the selected block. You can remove the filter by clicking the link that appears on the top of the table header row.

  2. A table that provides different performance metrics that are available under the following tabs of the table.
    IP Interfaces
    Shows all network interfaces that are part of the Ethernet and InfiniBand networks in the cluster. To view performance details in graphical format or to see that the events reported against the individual adapter, select the adapter in the table and then select View Details from the Actions menu.
    RDMA Interfaces
    Shows the details of the InfiniBand RDMA networks that are configured in the cluster. To view performance details in graphical format or to see that the events reported against the individual adapter, select the adapter in the table and then select View Details from the Actions menu.

    The system displays the RDMA Interfaces tab only if there are RDMA interfaces available.

    Networks
    Shows all networks in the cluster and provides information on network types, health status, and number of nodes and adapters that are part of the network.
    IP Addresses
    Lists all IP addresses that are configured in the cluster.

    To find networks or adapters with extreme values, you can sort the values that are displayed in the tables by different performance metrics. Click the performance metric in the table header to sort the data based on that metric. You can select the time range that determines the averaging of the values that are displayed in the table and the time range of the charts in the overview from the time range selector, which is placed in the upper right corner. The metrics in the table do not update automatically. The refresh button allows to refresh the table content with more recent data.

  3. A detailed view of performance aspects and events reported against each adapter. To access this view, select the adapter in the table and then select View Details from the Actions menu. The detailed view is available for both IP and RDMA interfaces.

The Network page depends on the InfiniBand sensor to display RDMA metrics. The InfiniBand sensor does not run if the IBM Storage Scale system is not installed by using the IBM Storage Scale install toolkit. In such cases, the RDMA graphs in the Network page shows "No Object". You can enable the InfiniBand sensors from the GUI by performing the following steps:

  1. Go to Services > GUI > Performance Monitoring.
  2. Select the InfiniBand sensor that is listed under the Sensors tab.
  3. Click Edit. The Edit Sensor InfiniBand dialog appears.
  4. Select the toggle button that is placed next to the Interval field.
  5. Select the frequency in which the sensor data is collected at the collector.
  6. Select the scope of the sensor data collection in the Node scope field.
  7. Click Save.

You can use the same steps to modify or start any performance sensors.