Ninguno

HOWTO: Collecting Exchange Data using Microsoft Performance Monitor
This HOWTO document details the step-by-step procedure on: how to gather counters for Microsoft Exchange; what they will be used for by EMC; how to upload them to the EMC FTP site; common issues during this procedure.

Steps to gather counters:
The following steps show how to gather the data from a single server – the stepsshould be

followed for every back-end (mailbox) server.
Step 1) In Perfmon.exe on a remote Windows 2003 or WindowsXP machine, select Counter Logs under Performance Logs and Alerts in the left pane and select New Log Settings as seen in Figure 1 below.

Figure 1. New Log Settings

Step 2) Select to Add Counters as seen in Figure 2 below.

Figure 2. Add Counters option

Step 3) Be sureto change the computer you are collecting data from to your Exchange server (don’t monitor your production server directly – use any other host) using the option Select counters from computer and then add the Performance objects listed in the email and be sure to select All counters and All instances for each object. Please create a text

file that contains which physical disks contain what data(i.e. “Physical disk 3 = E:\ = Exchange Storage Group 2 databases) – this will make analysis much more rapid. Upload this file when you subsequently upload the perfmon .blg. Add all counters and
all instances for the following objects:

• • • • • • • • •

Physical Disk – Avg Disk Queue Length Physical Disk – Avg Disk sec/Read Physical Disk – Avg Disk sec/Write Physical Disk – Disk bytes/secPhysical Disk – Disk Reads/sec Physical Disk – Disk Writes/sec MSExchangeIS – Active Connection Count MSExchangeIS – Active User count Processor - %Processor Time

For Physical Disk counters, don’t collect Totals, highlight the counters below, and highlight all of the partitions as shown below.

Step 4) Once complete you should see this summary screen with all of the Perfmon objects listed inthe email with an asterisk for each object for the Exchange server as listed below in Figure 4. Use this general tab to change the sample interval to 60 seconds.

Figure 4. Summary of all counters and instances for the Exchange server

Step 5) Select the Log files tab to change the file location and name of the output file if desired.

Figure 5. Changing the output file name and locationStep 6) Select the Schedule tab to create a start and stop schedule for data collection. In the screen shot perfmon is configured to start manually and to end after 2 hours, but configuration of an automated start/stop schedule is simple. For the purposes of analysis – a complete full 24 hr cycle is the minimum. Ideally, the counters should capture the peak period (generally Monday morning forExchange) as well as a full day of steady-state.

Figure 6. Configuring a log schedule

Step 7) Either start and stop the new counter log manually as seen in Figure 7 or retrieve the output file once the schedule is complete.

Figure 7. Starting the data collection manually

What information is being gathered from the counters:
The counters measured in the previous example are used to: 1.Rapidly scan overall server and storage subsystem health. Comprehensive Exchange/AD health checks via professional services engagements can provide an end-to-end assessment and resolution of any issues with the the Exchange environement and it’s dependencies 2. Highlight potential sources of performance issues 3. Gather information required for proper storage subsystem design, includingdetermining fit for iSCSI The table below higlights some of the expected values.
Counter
Processor: % Processor Time (_Total) Memory\Available Mbytes Memory\Free System Page Table Entries Memory\Pages/sec Memory\Pool Nonpaged Bytes Memory\Pool Paged Bytes

Expected Value
Should be less then 80% Minimum should be more than 50 MB Minimum should be more than 5000 Average should be less than 100, and the...