Hitachi

JP1 Version 12 JP1/Performance Management - Remote Monitor for Virtual Machine Description, User's Guide and Reference


1.9.6 Monitoring the network resources

This subsection explains how to monitor the network resources of a logical partitioning feature system.

Organization of this subsection

(1) Overview

In a virtual environment of a system with logical partitioning feature, multiple LPARs share the NIC of the host machine (including the machine built into the chassis). The NIC allocated to each LPAR is called the virtual NIC. The LPAR recognizes the virtual NIC as a typical NIC.

A server module is equipped with a built-in NIC, and connects to an external LAN via a switch module and management module built into the chassis. The built-in NIC equipped with the server module can be allocated by being shared or exclusively used by LPARs.

When PFM - RM for Virtual Machine is used, physical NICs equipped with the server module (built-in and external NICs) and virtual NICs associated with them are to be monitored as network resources.

By monitoring network performance data, you can keep track of the loads on these NICs, and thus you can take an appropriate corrective action.

The following two records are used to monitor the network resources. For details about each record, see 5. Records.

  1. PI_HNI record

    This record is used to monitor performance data of physical NICs on the host machine. By monitoring the amount of data traffic passing through the physical NICs, you can monitor the loads on the NICs. An overloaded NIC can affect the CPU utilization rate.

  2. PI_VNI record

    This record is used to monitor performance data of virtual NICs for LPARs. By monitoring the amount of data traffic passing through the allocated virtual NICs, you can monitor whether I/Os are concentrated on a specific virtual NIC.

The following figure shows the range of performance data collected in each record.

Figure 1‒105: Correspondence between records and data collection ranges

[Figure]

Tip

If an NIC is allocated in shared mode to a LPAR, CPUs on SYS2 process data, and therefore the amount of data traffic passing through the NIC affects the CPU load on SYS2. Thus, if the amount of data traffic passing through the NIC increases, CPU resource usage of SYS2 may increase excessively, causing performance deterioration in the LPARs. Note that if the NIC is allocated in dedicated mode to a LPAR, it does not affect the CPU load on SYS2.

(2) Monitoring examples

Using monitoring of the amounts of data traffic passing through the physical NICs on the host machine and passing through the virtual NICs for LPARs as examples, this subsection explains possible problems related to network resources, and how to solve those problems. The following figure shows the items monitored here and the flow of actions to take.

Figure 1‒106: Monitored items and flow of actions

[Figure]

(a) Example of monitoring the amount of data traffic passing through the physical NICs

You can estimate the load on a physical NIC from the data send/receive speed on that physical NIC. You can monitor the data send/receive speed on the physical NIC in the Rate field of the PI_HNI record.

The following figure shows an example of monitoring the data send/receive speed on the physical NICs.

Figure 1‒107: Example of monitoring the data send/receive speed on the physical NICs

[Figure]

Monitoring template report to be checked

Host Network Data

In this example, the traffic volume is low in the physical NIC1 and physical NIC3. On the other hand, the physical NIC2 remains overloaded. Take some actions, such as reviewing the configuration of logical partitioning feature or installing additional physical NICs.

(b) Example of monitoring the amount of data traffic passing through the virtual NICs

You can estimate the load on a virtual NIC from the data send/receive speed on that virtual NIC. You can monitor the data send/receive speed on the virtual NIC in the Rate field of the PI_VNI record.

The following figure shows an example of monitoring the data send/receive speed on the virtual NICs.

Figure 1‒108: Example of monitoring the data send/receive speed on the virtual NICs

[Figure]

Monitoring template report to be checked

VM Network Data

In this example, the traffic volume is low in the virtual NIC1 and virtual NIC3. On the other hand, the virtual NIC2 remains overloaded. In this case, the problem can be solved by reviewing how NICs are shared or exclusively used. If the problem persists, consider installing additional physical NICs or migrating the current environment to any other environment with logical partitioning feature.