Stare Down the Blue Screen of Death

boot

For those who haven't experienced it, BSoD is literally a blue screen displayed by Windows when a PC cannot recover from a system error. BSoD is always accompanied by arcane error messages and even more arcane instructions on what to do next. With Windows NT, 2000, and XP, BSoD occurs when the kernel encounters an error from which it cannot recover. This is typically caused by a driver that has performed what Windows calls an "illegal operation."

If the issue is with the OS—particularly the registry—the crash can be fairly easy to recover from. But deeper problems—especially those rooted in hardware issues such as the hard drive, RAM, or power supply—can be much more challenging. And these problems are compounded because critical data is often at risk, too.

Ingredients

Here are the components you'll need to get a system back up and running after a hard crash. Ideally, a system builder will have these components on hand beforehand, so they'll be ready whenever disaster strikes. Also, I'm assuming you're working with Windows XP Home on the system, or have a Professional Bootable CD on hand. Here's what you need:

• CCleaner: This free application is useful for cleaning up cache, temporary files, and other junk found on most hard drives. It also features a safe, conservative registry-cleaning function.
• Ultimate Boot CD: A bootable CD packed with utilities and applications to aid in diagnosing and repairing many system issues.
• Knoppix Live CD: This is a fully functional operating system that runs in a system’s memory and off a CD-ROM drive without needing to access the hard drive. It's useful for data recovery and troubleshooting when a hard drive isn’t working properly.
• CD Burner and Burning Software: You’ll need a CD burner, along with its accompanying software, to burn .iso files to create some of the boot disks I discuss in this Recipe.
• Memtest86: This is software designed to thoroughly test a system’s RAM chips.
• PC Inspector File Recovery: Free and powerful software used for data recovery.
• SpinRite 6 (optional): You can use this for checking, fixing, and maintaining the hard drive's health.

I'll discuss these packages in more detail, and provide links where you can download them, in the rest of this Recipe.

Common-Sense Precautions Before a Crash

The first step toward restoring a system should be taken before the system crashes. Namely, making backups. Sure, your customers have heard this a hundred times before. But backing up all critical data on a regular basis is a must. These days, users don't need sophisticated automated application suites, either. As long as they have a CD/DVD burner or small USB portable storage "thumb" drive, they're set. Instruct them to make two copies of all critical data: one for themselves, and one to store someplace safe.

Assuming your customers have been fairly diligent about making backups, here are a few other tips to protect up-and-running systems against possible future disasters:

• Create a list of all applications and utilities on their system that are no longer used. No one needs an installation of TurboTax from three years ago, do they? Once you have your list, deploy the Windows Control Panel Add or Remove Programs function to uninstall the clutter.
• Defrag hard disks regularly. The Disk Defragmenter included with Windows is adequate, although it cannot be scheduled. To use this Windows tool, right-click on the hard drive in My Computer, then click on Tools > Properties.
• Clean up old temp files and cache. Windows' built-in Disk Cleanup function does a decent job. It can be found by opening My Computer, right-clicking on the hard disk and selecting Properties > Disk Cleanup on the General tab.
• In addition, specialized applications can clean out a lot of additional junk. A terrific disk cleanup application is CCleaner. CC is short for "crap cleaner," and this application's primary purpose is to clean out extraneous stuff like the Internet Explorer cache, cookies, and unnecessary temp files. Adding to its appeal is a conservative registry-cleaning function.
• As a bonus, CCleaner can help clean the Windows Registry. Cleaning the Registry manually is tricky business, even when you know what you’re doing. But CCleaner saves you from that headache, and it does a nice job automatically. The program wisely prompts you to create a backup copy of the registry. I've been using it for more than a year without any issues. Best of all, it's free.
• First Aid for Minor Repairs
• But what if you're facing a system that simply won't boot, and/or keeps throwing a Blue Screen of Death. What to do? Here's how to get started restoring the system.
• If the system shows the BSoD, the information on the screen is typically too arcane to be of much use. Your best option is to instead perform a reboot. First, power down the machine by hitting the power button. Then, give the machine 15 or 20 seconds to sit. Then power it back up. If the system boots up successfully, take a deep breath, smile, and back up the data as soon as you can.
• If, in the course of a standard startup, the system fails to successfully boot into XP, power the system down and try again. Its next reboot will present you with three options: Boot using Safe Mode; boot using Last Known Good Configuration; and boot normally. I recommend that you try the Last Known Good Configuration recovery option. This essentially uses a snapshot of your system’s configuration the last time it successfully booted into Windows. If that doesn’t work, tap the F8 key while you reboot yet again; this will bring you to the Safe Mode menu. This menu includes options for both Last Known Good Configuration as well as several versions of Safe Mode.
• The various Safe Mode options boot the system with just a skeleton crew of device drivers and services. Since basic Safe mode doesn’t support networking, if you need connectivity, select Safe Mode with Networking. Safe Mode is particularly useful for troubleshooting issues that may have been caused by newly installed hardware or software—or by spyware, adware, or viruses. By booting into XP using Safe Mode without networking, you can also take the time to run anti-virus and anti-spyware utilities, and reduce the risk of infecting others on a network or even re-infecting yourself!
• Rebooting with the Last Known Good Configuration will attempt to boot your system with the last settings that worked, essentially turning back the clock. But this is far from a guarantee of smooth sailing. The issues that caused the system to fail to boot are quite possibly still lurking, and you may find you need to repair the OS. Let's move to that phase now.
• XP Installation Repair with Recovery Console
• If an installation of Windows XP is detected on the hard drive when you start from a Windows XP install disc, you’ll be prompted to either perform a completely fresh install over the top of the old OS, or attempt to repair the existing installation.
• Choosing the Repair option will bring you to into the Recovery Console. You'll need an administrator password to continue; the Recovery Console is very powerful and allows full access to the drive.
• The Recovery Console will be familiar to anyone who used to work with the good old DOS command line. It supports moving, copying and renaming files; formatting the hard drive; and even re-installing critical OS files, such as NTLDR or NTDETECT.COM (without which XP won’t boot). You can also fix a Master Boot Record gone bad or repair the file system boot sector. Powerful stuff.
• There's a very helpful article on Microsoft's support site that gives an overview of this function: How to install and use the Recovery Console in Windows XP. This article also offers a link to download a version of the Recovery Console that fits on six floppy disks; you'll need this if you don't have access to the system's original XP disc. Also, a good in-depth explanation of the Recovery Console in WinXP can be found on this Kelly's Korner page. Major Overhauls
• In this section, I'll take you through the biggest chores of getting a crashed system back up: working with boot disks, and salvaging critical data with recovery software.
• A boot disk is generally a slimmed-down version of an OS that allows you to boot directly from either a CD-ROM or floppy drive—that is, without the hard drive. To use a boot disk, you’ll need to enter the BIOS while the system first powers up by hitting F2.
• Depending on the system's BIOS, the hard drive is usually set by default as the first place the system looks for an OS. Instead, depending on the type of boot disk you’re using, set either the floppy drive or CD-ROM drive as the first item in the boot sequence. Next, save your changes to the BIOS. Then reboot. Assuming you aren’t facing catastrophic hardware issues, the system will look for an OS in the CD or floppy drive, and will then boot it up.
• While a bootable copy of the Windows XP CD is helpful, it isn't the only option for getting a system back up. There are dozens of customized boot disks available, ranging from Microsoft's own recommendations to a number of proven Linux distributions. Here's a few of these better options for Windows and Linux:
  • For starters, read this helpful article from Microsoft's support site: How to obtain Windows XP Setup boot disks. This article also contains links to Windows XP Setup boot disks, so you can run Setup without a bootable XP disc.
  • A number of bootable operating systems are available as "live" CDs, as well. A live CD is a complete OS environment that doesn't use the hard drives at all. Instead, it runs completely off the CD or DVD drive and in RAM. Most live CDs tend to be flavors of Linux. Among the most popular are Knoppix (which I discuss below), and Damn Small Linux--a fully functional OS that takes up less than 50 MB.
  • For those interested in trying one of the dozens of various boot disk flavors (for a variety of operating systems), Bootdisk.com is a worth exploring.
  • A respected live CD for the Linux operating system is the Knoppix Linux distribution. Knoppix Live can be used to boot a system with a damaged OS, and it may be able to mount and read data from damaged FAT32 and NTFS-formatted hard drives. Also, Knoppix recognizes certain removable media, such as USB thumb drives; this lets you recover data by moving it from a damaged disk. I can also recommend this step-by-step tutorial: Computer First Aid Using Knoppix.
  • The Ultimate Boot CD comes in two versions, both of them free. The Basic version lets you boot a system from the CD and run dozens of utilities, including anti-virus, file-recovery, and hard-disk maintenance. The Full version adds a version of the Knoppix Linux distribution mentioned above.
• While boot disks and live CDs can be helpful, some have a steep learning curve. I recommend that you first burning a few of them and then take them for a test drive now, while your clients' systems are still working correctly. You don't want to run experiments in the middle of a disaster.
• Salvaging Critical Data
• While sending a damaged hard drive to a professional data-recovery service is an option, it's likely to cost thousands of dollars. That may be an acceptable last resort for users with unique, impossible-to-replace data. But here are a few options that can help you recover data quickly for a lot less money:
• PC Inspector File Recovery is freeware from Convar Deutschland that can recognize the partitions created on each hard drive and recover data that's been deleted or damaged. Unlike some applications that need to be installed beforehand, File Recovery doesn't need to be running in the background to recover lost or deleted data. In my own tests, I've found PC Inspector to be effective, though not foolproof: It successfully recovered about 90 percent of my deleted files.
• Another well-regarded option is SpinRite 6 from Gibson Research. This tool can test a hard drive's health as a preventative measure, and then recover bad data once a crash has occurred. Even better, SpinRite 6 costs just under $30, making it a cost-effective alternative to expensive data-recovery specialists. Hardware Troubleshooting
• System crashes are most commonly are caused by hard drives, RAM, and power supplies. Before you replace any hardware or drive yourself crazy troubleshooting, give the system a once-over: Make sure all connections to all drives are tight, the power supply is firmly plugged in, and all cards and RAM boards are seated properly. Sounds obvious, but these are common problems.
• If, after checking these connections, the system still won't boot, here are some other hardware options:
• Hard Drives: Identifying problems on a hard drive before they happen involves a bit of voodoo. Luckily, most major manufacturers offer hard-drive diagnostic tools on their sites, and that's a good place to start. For another option, TACHtech Corp. offers a sprawling community site that includes a list of common manufacturers and their respective hard-drive utilities. Yet another option is to use SpinRite (discussed above) to run the hard drives through a series of checks to identify and correct issues.
• If you suspect (or the disk check has confirmed) that the hard drive has bad sectors, then get all data off the drive as quickly as possible and install a new drive. While modern drives can work around bad sectors, new hard drives with enormous capacity are now so cheap, there's no excuse for not replacing a questionable disk.
• RAM: The quickest way to check to see if a system's memory is working properly is to right click on My Computer, select Properties, and then click on the General tab. If you’ve confirmed that the RAM is seated correctly, but the amount of memory displayed doesn’t match up to what you believe you have installed, a more thorough test is needed.
• One useful (and free) tool for doing that is the Memtest86, which can be downloaded and burnt as a bootable disk. This will do a very thorough check of the RAM in the system and will return pages of extremely detailed logs. Although Memtest86 may be overkill for basic troubleshooting, it’s certainly informative.
• Power Supplies: Both sporadic crashing and a complete failure to power up can be caused by a power supply that's either going bad or is simply overtaxed. With power-hungry motherboards, drives, and audio/video cards becoming the norm, this is becoming a common problem. In general, you should be safe replacing any old power supply with a new one that supplies 350 or 400 watts. But check first with the manufacturers of the components you’re using for specific guidelines.
• Full OS Reinstall Shortcuts
• A full OS reinstall tends to be a last resort, although some users swear by a fresh install a couple of times a year! Getting the environment just right—with all applications, settings, and data back the way your customer wants—is the time-consuming part.
• One alternative to starting completely fresh with a reinstall is using an application such as Norton Ghost to essentially create a clone of the hard drive. Though Ghost is overkill for people with just a computer or two at the home or office, it's worthwhile for system builders who are building many similar or identical systems. Ghost retails for about $70.
• Another shortcut: Pull some of the system's configuration settings from an old restore point on the machine. For help with this, Digtalwebcast.com offers this tutorial, Windows XP Crashed? Here's Help. This article shows how to retain as much of the customized user attributes (profiles, applications, etc.) as possible. Be aware that this might not be the correct solution for you—especially if you suspect the system's Registry may be at the root of the problem. But the article might also save you some headaches the next time you want to start fresh with a clean install of XP.
• If you followed the steps and advice in this Recipe, and you've visited the links and read through the articles I've recommended, you need never again be intimidated by the dreaded Blue Screen of Death.
• PETER HAGOPIAN is a Philadelphia-based freelance writer who's been covering technology and music for more than a decade.