Wednesday, November 25, 2009

Computer crashes and whatnot.

Some of you may recall some months ago when my Vista computer was struck dead by the dreaded "Black Screen of Death" (KSOD to you technically inclined). If you Google around the Internet, you'll see that KSOD seems to be a relatively common failure, thought to be caused by a number of different issues but all of which result in a catatonic, non-functioning computer.

Like most things in life, it felt worse while it was happening than reflecting on it now, and in the end, I was able to fix it and even got my system into a better state than it was previously, being forced to stop my normal hectic life in order to recover, tidy up and get rid of dunnage in my computer.

Unfortunately, KSOD struck again this past weekend, and even the otherwise wonderful Windows 7 was not immune. It happened after what I thought was a rather innocent "chkdsk /f" to fix errors on my hard drive but resulted in it deleting perfectly good permissions in the NTFS file system access control list and leaving me with a dead computer.

Only this time, I don't have a good backup... more on this later. Ironically, I ran the disk checker in an attempt to fix my backup problem.

So once again, I am forced to stop the busy planning, running around, doing and yes, procrastinating in order to put back my computer into a working state.

Taking a step back to reflect on the situation, however, I can make two related observations. The first is how much of all the things I consider to be important in my life resides in that little digital homestead that is my computer. And the second is how much of my own identity and happiness I've permitted to be tied up to these material things.

Anyway, my new hard drive should be arriving by post anytime now and then I can start the lengthy process of reinstalling and reconfiguring Win7 and all my applications. Luckily, since KSOD results in a non-booting computer but doesn't actually delete data, ultimately I will not have lost much except for a lot of time and perhaps some pride.

It's also a good opportunity for me to install the 64-bit version of Win7, which I chose not to do previously coming from a 32-bit Vista system because I would have to reinstall all my programs. I guess I have time to do that now.

So anyway, some notes for people who have encountered KSOD and are frantically searching the Internet for solutions...

On KSOD causes and solutions:
A lot can be found on Google with the search terms "KSOD fix chkdsk". There seem to be a number of different causes including corruption of RPC server permissions, registry permissions, Event Log permissions and chkdsk-related permission corruption. This is a good page that describes a number of them with possible solutions:

Both times in my case (Vista and Win7), I believe that the KSOD was caused by chkdsk going haywire. In the Vista case, my computer had blue screened and crashed, and after restarting the system, Vista ran the mandatory chkdsk because "Windows wasn't shut down properly". In the Win7 case, I started "chkdsk /f" myself. Both resulted in chkdsk "Replacing invalid security id with default security id for file " and corrupting the ACL. Both cases ended up with KSOD upon reboot.

Microsoft has a hotfix for similar symptoms in Windows Server 2003, but not for Vista or Win7:

I don't know if this is the same issue, but it doesn't specifically mention Vista or Win7 and probably isn't compatible.

Althought the resetting file permissions post near the end of the KSOD thread on matched my symptoms and seemed promising, it didn't work for me and I couldn't get past the KSOD.

Also, because the file permissions are messed up, I still couldn't get access to the files after putting the hard drive in a USB enclosure and connecting it to a Vista system. I got arond this by booting into UBCD4Win/BartPE, which runs a modified XP system and doesn't enforce ACL. I suppose you could get Vista to read the hard drive by turning off UAC or trying to Take Ownership of the files, although I didn't try this myself.

On hard drives:
I ended up buying two drives - a 1TB Seagate FreeAgent Desk external USB hard drive, and a 500GB Western Digital Scorpio Blue laptop hard drive.

The FreeAgent drive is fantastic. It seems reliable so far and the industrial design is beautiful. I love how the white LED pulses through the perforated metal cover when the drive is in use. Stunning! And the power adapter is small and comes with both exchangable UK and EU plugs.

I bought the WD Scorpio Blue hard drive because it was available quickly and seems to be the default workhorse model that most people buy - dependable and robust.

Originally, I wanted to get a drive with a shock sensor built in, like the Seagate Momentus 5400.6 (ST9500325ASG), the Momentus 7200.4 (ST9500420ASG) or the WD Scorpio Black (WD3200BJKT). Seagate drives with the G-Force sensor end in "ASG" and WD drives with the freefall sensor end in "BJKT".

Long story short - nobody seemed to have any of the sensor-enabled drives in stock except the 7200.4, but a quick Google of "Seagate 7200.4 click problem" reveal that this drive might have a serious design fault.

On Backup solutions:
I have used both Aronis TrueImage and Rebit backup solutions.

While Acronis offers more features and capabilities, I never could discipline myself to make regular backups often enough to matter. Also, only the 2010 version supports Win7, although if you search the Acronis forums, it appears that Win7 users are having problems with restoring backup archives.

Rebit is meant to turn a USB hard drive into a backup appliance that you plug in to your computer and forget about it, like Time Machine for OS X. However, I never could get it to work right, although the second copy I bought for a friend works brilliantly for them. One suggested solution by Rebit to troubleshoot the system is to perform a chkdsk /f on the host drive...

Anyway, this post is much longer thatn I wanted, but I hope my experiences help those of you who found this page by searching for KSOD. In the end, none of the "quick" solutions worked for me and I am reinstalling my system and working on having better backups.

Good luck! Your data probably is fine and with a little peserverance, you'll be back up and running in no time. Take care.

No comments: