Windows Vista Improvements for Hardware and Driver Troubleshooting

Windows Vista includes several improvements and new features that will simplify how you troubleshoot hardware problems, allowing you to reduce client computer downtime. The following sections describe these improvements.

Windows Memory Diagnostics

Application failures, operating system faults, and Stop errors are often caused by failing memory. Failing memory chips return different data than the operating system originally stored. Failing memory can be difficult to identify: Problems can be intermittent and might only occur under very rare circumstances. For example, a memory chip might function perfectly when tested in a controlled environment but begin to fail when used within a hot computer. Failing memory can also cause secondary problems, such as corrupted files. Often, administrators take drastic steps to repair the problem, such as reinstalling applications or the operating system, only to have the failures persist.

Windows Vista includes Windows Memory Diagnostics to help administrators track down problems with unreliable memory. Previously, this technology was only available as a download and required installing the tool on a bootable floppy disk. In Windows Vista, if Windows Error Reporting (WER) or Microsoft Online Crash Analysis (MOCA) determine that failing memory might be the cause of an error, the software can prompt the user to perform memory diagnostics without requiring an additional download or separate boot disk. Additionally, you can run Windows Memory Diagnostics by choosing a special boot menu option or by loading the Startup Recovery Tools from the Windows Vista DVD.

If memory diagnostics identify a memory problem, Windows Vista can avoid using the affected portion of physical memory so that the operating system can start successfully and avoid application crashes. Upon startup, Windows Vista provides an easy-to-understand report detailing the problem and instructing the user on how to have the memory replaced.

Disk Failure Diagnostics

Disk reliability problems can vary in severity. Minor problems can cause seemingly random application failures. For example, if a user connects a new camera and the operating system fails to load the driver, disk corruption may be causing the problem. More severe problems can result in the total loss of data stored on the hard disk.

Windows Vista can eliminate much of the impact of a disk failure by detecting disk problems proactively, before total failure occurs. Hard disks often show warning signs before failure, but earlier Windows operating systems did not record the warning signs. Windows Vista listens for evidence that a hard disk is beginning to fail and warns the user or the support center of the problem. IT can then back up the data and replace the hard disk before the problem becomes an emergency. For administrators, Windows Vista acts as a guide through the process of backing up their data so that they can replace the drive without data loss.

Most new hard disks include Self-Monitoring Analysis and Reporting Technology (SMART) and Disk Self Tests (DSTs). SMART monitors the health of the disk using a set of degradable attributes, such as head-flying height and bad block reallocation count. DSTs actively check for failures by performing read, write, and servo tests.

Windows Vista queries for SMART status on an hourly basis and regularly schedules DSTs. If Windows Vista detects impending disk failure, Windows Vista can launch disk diagnostics to guide the user or IT professionals through the process of backing up the data and replacing the disk before total failure occurs. Windows Vista can also detect problems related to a dirty or scratched CD or DVD, and instruct the user to clean the media.

You can configure disk diagnostics using two Group Policy settings. Both are located in Computer Configuration\Administrative Templates\System\Troubleshooting And Diagnostics\Disk Diagnostic.

• Disk Diagnostic: Configure Execution Level Use this policy to enable or disable disk diagnostic warnings. Disabling this policy does not disable disk diagnostics; it simply blocks disk diagnostics from displaying a message to the user and taking any corrective action. If you have configured a monitoring infrastructure to collect disk diagnostic events recorded to the Event Log, and you prefer to manually respond to events, you can disable this policy.

• Disk Diagnostic: Configure Custom Alert Text Enable this property to define custom alert text (up to 512 characters) in the disk diagnostic message that appears when a disk reports a SMART fault.

For disk diagnostics to work, the Diagnostic Policy Service must be running. Note that disk diagnostics cannot detect all impending failures. Additionally, because SMART attribute definitions are vendor-specific, different vendor implementations can vary. SMART will not function if hard disks are attached to a hardware RAID controller.

Self-Healing NTFS

Windows Vista now includes self-healing NTFS, which can detect and repair file system corruption while the operating system is running. In most cases, Windows Vista will repair file corruption without disrupting the user. Essentially, self-healing NTFS functions similar to Chkdsk (described later in this chapter), but happens in the background, without locking an entire volume. Specifically, if Windows Vista detects corrupted metadata on the file system, it invokes NTFS’s self-healing capabilities to rebuild the metadata. Some data may still be lost, but Windows Vista can limit the damage and repair the problem without taking the entire system offline for a lengthy check–and-repair cycle.

Self-healing NTFS is enabled by default and requires no management. Instead, it will serve to reduce the number of disk-related problems that require administrative intervention. If self-healing fails, the volume will be marked dirty, and Windows Vista will run Chkdsk on the next startup.

Reliability Monitor

Two of the biggest challenges of troubleshooting hardware problems are determining when the problem began occurring and what might have changed on the computer to introduce the problem. Windows Vista provides the Reliability Monitor snap-in (part of the Computer Management console) so that you can easily view application installations, driver installations, and significant failures over several weeks or months. Figure 1 shows Reliability Monitor.

Figure 1. Reliability Monitor provides you with a history of changes and problems.

Reliability Monitor gives you a day-by-day analysis of application and driver installations, application failures, hardware failures, operating system failures, and other uncategorized failures. Reliability Monitor will be extremely valuable when troubleshooting hardware problems reported by end users because it allows you to quickly assess changes even if the user is not aware of them.

Improved Driver Reliability

Windows Vista drivers should be more reliable in Windows Vista than they are in previous versions of Windows. Improved input/output (I/O) cancellation support has been built in to Windows Vista to enable drivers that might become blocked when attempting to perform I/O to gracefully recovery. Windows Vista also has new Application Programming Interfaces (APIs) to allow applications to cancel I/O operations, such as opening a file.

To help developers create more stable drivers, Microsoft provides Driver Verifier. Developers can use the Driver Verifier to verify that their drivers remain responsive and to ensure that they correctly support I/O cancellation. Because driver hangs can affect multiple applications or the entire operating system, these improvements will have a significant impact on Windows Vista stability. This improvement requires no effort from administrators; you will simply benefit from a more reliable operating system.

Improved Error Reporting

Windows Vista will offer improved application reliability, and the new error reporting capabilities will allow applications to continue to become more reliable over time. In earlier versions of Windows, application hangs were very hard for developers to troubleshoot because error reporting provided limited or no information about hangs. Windows Vista improves error reporting to give developers the information they need to permanently resolve the root cause of the problems, thus providing continuous improvements in reliability.