NFS Read Cache Test

This test monitors the read cache of the NFS datastore and reports the level of I/O activity on the NFS datastore so that administrators are notified of the issues related to the datastore well before they impact the performance of the NFS datastore.

Target of the test : An Atlantis ILIO server

Agent deploying the test : An internal agent

Outputs of the test : One set of results for the Atlantis ILIO server that is to be monitored.

Configurable parameters for the test
Parameter Description

Test Period

How often should the test be executed.

Host

The IP address of the host for which this test is to be configured.

Port

The port at which the specified Host listens. By default, this is NULL.

Measurements made by the test
Measurement Description Measurement Unit Interpretation

Cache hits

Indicates the number of requests served from the NFS read cache.

Number

Ideally, the value of this measure should be high. A consistent drop in this value indicates that read I/O requests are directly served from the backend storage which is a cause of concern.

Cache misses

Indicates the number of requests that were not served from the NFS read cache.

Number

A low value is desired for this measure. A high value indicates that the cache is used ineffectively meaning that the cache is unable to serve most I/O requests thus increasing the load on the backend storage.

Cache hit ratio

Indicates the percentage of requests served from the NFS read cache.

Percent

A value of 80% and above for this measure is a sign of a healthy cache. A very low value indicates that the cache is unable to serve most of the I/O requests and the data is served directly from the backend storage.

Hits bypassed

Indicates the number of requests that bypassed the cache and were served directly from the backend storage.

Number

A low value is desired for this measure. A high value may indicate that the cache is not used or the requests sent to the cache bypasses without hitting the cache, and accesses the backend storage directly.