eG Monitoring
 

Measures reported by IBMV7000CPUTest

A Node canister is a hardware unit that includes the node hardware, fabric and service interfaces, and serial-attached SCSI (SAS) expansion ports. Each control enclosure of the IBM Storwize v7000 system consists of two node canisters. In a clustered system therefore, where a collection of control enclosures are managed as a single system, multiple node canisters are bound to exist. At the heart of each node canister is an Intel Jasper Forrest (Sandy Bridge) based quad core CPU. Optimal usage of this CPU is key to reducing latencies in I/O processing. In other words, if a node canister utilizes its processing power excessively, it could cause I/O processing delays, which in turn can scar the user experience with the storage system. This is why, administrators should keep an eye on the CPU usage of each canister, so that abnormal usage patters can be detected proactively and treated, before end-users complain. This is where the IBMV7000CPUTest test helps.

This test monitors the CPU usage of each node canister of the IBM Storwize v7000 storage system and reports how long every node canister used its CPU. In the process, the test points to those canisters that are over-utilizing their processing power.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Cpu_busy Indicates the time for which the CPU of this node canister was busy. Millisec A high value for this measure indicates that the CPU is busy for a longer time and the processor is working overtime to complete the job.
Cpu_busy_percent Indicates the percentage of time the CPU of this node canister was busy. Percent A value close to 100 indicates processing bottleneck.