How to monitor disk IO issues in Windows
search cancel

How to monitor disk IO issues in Windows


Article ID: 135827


Updated On:


CA Identity Manager CA Identity Governance CA Identity Portal CA Identity Suite


CA Identity Manager like other applications relies on data stores for example writing database transactions or replicating data for clustering.  In some circumstances latency may be encountered as a result of too much read \ write activity from competing applications.  Microsoft Windows operating system provides a performance monitoring application which can be useful in identifying any specific issues.


Release : 14.x

Component : Identity Minder(Identity Manager)


To add Perfmon counters to monitor Disk IO (Input\Output) Activity please execute the following steps.

1. Run Performance Monitor (Perfmon.exe) this is part of the Windows Administrative tools.

2. Navigate to Performance -> Data Collector Sets -> User Defined

3. In the Right hand Pane right click and Select New - > Data Collector Set

4. Provide a Friendly name (i.e. Task Set)

5. Click on "Create Manually (Advanced)"

6. Click "Next"

7. Under "Create Data Logs" select Performance Counter

8. Click Finish

The data set (i.e. Task Set) should now appear in the right hand pane.

1. Double click on the data set (ie Task Set)

2. "DataCollector01" should appear in the right hand pane.  Double click on the "DataCollector01"

3. In the "Performance Counters" tab click Add.  

4.  A list of the available counters appears on the screen,  Navigate to "LogicalDisk" and click to highlight (do NOT expand the list).  Under instances of Selected objects select the specific volume (drive) where the journal files are located (ie C:) and click "Add".

This will list all LogicalDisk performance counters associated with activity on the specified volume (drive).  You will see something like the following in the Added Counters window:

5. Click OK to return to the "DataCollector01" Properties screen.

6. Leave the other values at default settings - Log Format set to "Binary" and Sample Interval set to "15" Seconds and Click "OK"

This creates the basic dataset.  To scheule the Perfmon.

1. The data set (ie Task Set) should now appear in the left hand pane.  Right click and select "Properties"

2. The DataSet Properties includes a number of tabs.  

(a) Use the General tab to set user credentials with the relevant privileges to run the perfmon (By default this is set to SYSTEM)

(b) Use the Directory tab to set the data path of where you would like to store the logs files (by default this is set to the %systemdrive% - ie C:\) we would recoomend that the log files are stored on a drive which is not being monitored as this will skew the statistics.

(c) Use the schedule tab to Set the days to run and the time to commence collection.  This would be specific to the customers working practices, for example if the issue under investigation is seen commonly at the start of normal office hours the schedule should be set for Monday - Friday commencing at 09:00am

(d) Use the Stop Condition to set the parameters to stop (ie the Overall Duration of monitoring is 8 hours).

3. Click OK when all required parameters are configured.

4. You can start data collection by clicking on the DataSet (ie Task Set) right click and press start.

Additional Information

All "LogicalDisk" perfmon metrics are useful for analysis but the following are particularly interesting.

Current disk queue length:

Current Disk Queue Length is the number of requests outstanding on the disk at the time the performance data is collected. This means that the disk is not able to honor I/O requests as fast as they are being made. If the disk has a consistently high value for the Current Queue Length counter over multiple consecutive samples this suggests I/O requests latency is on the increase.

Disk Averages (Idle\Read\Write):

This will give us a snapshot of the IO Trends over the period of the sample

Disk Read\ Write\Transfers:

This will give us the granularity of the IO activity