WMI Support for SMART Drives

Updated: January 13, 2003

Manageability and predictive failure capability are provided through the Microsoft Windows Management Instrumentation (WMI) capabilities in Windows 2000, Windows XP, and Windows Server 2003. Self-Monitoring, Analysis, and Reporting Technology (SMART) hard drive support was enhanced with the addition of four GUIDs that are surfaced into the root\WMI schema. These GUIDs are:

• Read Failure Predict Status: MSStorageDriver_FailurePredictStatus

• Read Failure Predict Data: MSStorageDriver_FailurePredictData

  Notice that all such data is vendor specific.

• Failure Predict Event: MSStorageDriver_FailurePredictEvent

• Perform Failure Predict Function: MSStorageDriver_FailurePredictFunction

  Notice that this GUID has five methods associated with it:

  • Enable/Disable Hardware Failure Prediction

  • Enable/Disable Failure Prediction Polling

  • Enable/Disable Performance Degradation for Better Prediction

  • Get Failure Prediction Mechanism, such as SMART IOCTL, SCSI SMART, and IOCTL for filter drivers

  • Enable Offline Diagnostics; IDE SMART only

Management applications can take advantage of these capabilities in several ways. Examples include taking an action based on seeing the "Failure Predict Event" launch or periodically polling the condition of "Read Failure Predict Status."

One of the advantages of these enhancements to WMI is that user-mode applications no longer must be polled to be notified of a change in state. Because there will be an event mechanism available, polling becomes an optional method to monitor change in state. Other advantages include the unification of IDE and SCSI SMART, plus the fact that placing this capability in the storage stack instead of only in disk stacks allows the possibility for this type of information to be leveraged with respect to tape drives, CD-ROM drives, and so on.

By default, WMI polls the hard disk periodically; this polling frequency can be configured using a WMI (WBEM) method call, for example, ExecMethod or ExecMethodAsync.

For IDE, this polling consists of sending the READ_SMART_STATUS command. If this status indicates an error, a WMI event will be launched that includes the SMART attribute data. There are also WMI (WBEM) classes that will cause the SMART status and attribute data to be returned when queried.

For SCSI, the storage stack monitors for the occurrence of a CHECK CONDITION with a DRIVE FAILURE PREDICTION sense code, at which point a WMI event will be launched. This is polled for by a periodic read of block 0 of the SCSI disk being monitored; this polling frequency is also configurable using ExecMethod or ExecMethodAsync.

The mechanism used to provide SMART data using WMI has also been designed with extensibility in mind; drive manufacturers will be able to enhance or replace the SMART reporting mechanism in drives which use alternative mechanisms for predictive failure analysis.

Call to action for WMI and SMART drives:

• For complete details about the GUIDs and parameters for methods described in this article, see the WMI documentation in the Microsoft Platform SDK and the Windows DDK.

 Top of page