Store Interfaces Test

The core data storage repository for Microsoft Exchange server 2007/2010 is the Microsoft Exchange Information Store service. This test is useful in isolating and determining issues involving the interfaces between the Microsoft Exchange Information Store service on the Mailbox server and Edge/Hub Transport servers. Unlike Exchange Server 2003, Exchange 2007/2010 communicates with Hub Transport servers via RPC, not Simple Mail Transfer Protocol (SMTP), and therefore latency and queuing are a greater concern. This test isolates and determines issues involving the interface between the Microsoft Exchange Information Store service on the Mailbox server and Hub Transport servers.

Target of the test : A server configured with the Hub Transport role

Agent deploying the test : An internal agent

Outputs of the test : One set of results for each interface between the Exchange store and the Hub transport server being monitored.

Configurable parameters for the test
Parameters Description

Test Period

How often should the test be executed.

Host

Indicates the IP address of the Hub Transport server.

Port

The port number of the Hub Transport server. By default, this is 691.

Measurements made by the test
Measurement Description Measurement Unit Interpretation

Active connections to connection caches

Indicates the number of active connections in all connection caches for this interface.

Number

 

Idle connections to connection caches

Indicates the current number of idle connections in all connection caches for this interface.

Number

Ideally, the value of this measure should be zero, as idle connections in a cache are only resource drainers.

Current RPC requests outstanding

Indicates the current number of outstanding RPC requests for this interface.

Number

Ideally, the value of this measure should be 0.

Data transfer over an RPC call - average

Indicates the average number of bytes, sent to the server in one RPC call via this interface.

KB

These measures serve as good indicators of the data load generated by the RPC traffic to and from the Hub transport server.

Avg. data received per RPC call

Indicates the average number of bytes, received from the server in one succeeded RPC call via this interface.

KB

Average RPC latency

Indicates the average latency in milliseconds, averaged across all RPC operations in the past 1024 RPC packets.

Msecs

Average is calculated over all RPCs since exrpc32 was loaded.

Ideally, this value should be less than 25 ms.

High RPC latencies often cause significant delays in the Mailbox server – Hub Transport server interactions, and hence need to be eliminated.

Slow RPC requests

Indicates the percentage of slow RPC requests in the RPC queue, currently.

Percent

A slow RPC request is one that has taken more than 2 seconds. Any value higher than 5% for this measure is a cause for concern.

 

 

RPC requests failed

Indicates the percentage of requests in the RPC queue that currently failed.

Percent

Ideally, this value should be 0. A non-zero value for this measure migt warrant an investigation.

RPC requests succeeded

Indicates the percentage of RPC requests in the queue that currently succeeded.

Percent