Good that I’ve seen this after I setup Vista on my SBS network: "Looking to Run Vista RC1 on SBS 2003?"
When I was running Vista in Virtual Server through remote desktop, I watched Explorer crawl and die. Then Inoted a blue screen showing in the thumbnail.
By the time I got to the remote console to see what was happening, Vista had rebooted already, and the event viewer and dialog on login said, yes, it bluescreened!
From the dialog:
Problem Event Name: BlueScreen
OS Version: 6.0.55126.96.36.199.256.1
Locale ID: 1033
Additional information about the problem:
OS Version: 6_0_5536
Service Pack: 0_0
Files that help describe the problem:
The entry in the event viewer had an interesting item. Note the fault bucket line:
Log Name: Application
Source: Windows Error Reporting
Date: 8/31/2006 1:20:39 PM
Event ID: 1001
Task Category: None
Fault bucket 0x50_nt!MiMakeSystemAddressValidPfn+26, type 0
Event Name: BlueScreen
Running the debugger and doing "!analyze -v" didn’t help; Vista symbols don’t seem to be in the main symbol repository yet, though as you can see from the fault bucket info, Vista gets its symbols from somewhere.
It’s a bit (read: a lot) technical but perhaps I can put that fault bucket info into a search engine without running the debug tools like I do now.
Anyway, the error (0×00000050 PAGE_FAULT_IN_NONPAGED_AREA) suggests corrupt hardware, according to Microsoft. I’m not sure how this applies since I’m running Vista in a VM, but I did scan the virtual drive from within Vista for errors.
In the meantime, wouldn’t this dialog be better? I think so!
I’m posting this from my copy of Vista build 5536 (pre-RC1) running inside Virtual Server. I ran into a problem getting Vista to install on Virtual PC. When I tried to mount the ISO image, VPC gave me: "This File does not appear to be a CD Image
CD Image files must be greater than 2MB and a multiple of 2K in size"
Starting the Vista install from the physical disk (attached to the VM) worked, but thie install barfed with Vista reporting that INSTALL.WIM–Vista’s base installation image–was corrupted. I burned another DVD on a different machine with different burning software–same result. I checked the MD5 sum against the one listed on the pre-rc1 download page and it matched.
I then tried installing from a VM in Virtual Server on another machine I have and it worked; I then transferred my VHD image back to the workstation I use for testing and ran it under Virtual Server. Worked, after I reset the network configuration.
After installing the virtual machine add-on for Vista Beta 2, I have a reasonably responsive, though plain, client. (no Aero since I’m running from the Terminal Services client).
BTW, if you are running Windows in a VM to use as a desktop, I strongly recommend enabling remote access so you can access it from Remote Desktop. When you run Remote Desktop (Start/Run/mstsc.exe) open the Options dialog and under Sound, specify "Bring to this computer". That’s the only way you can get sound when you run Vista as a VM.
More to follow.
In my previous post, I talked about the bad caps I found in my personal PC’s motherboard. Sadly, I think I’ve found another bad cap, this one in a 5-year old Dell 1500SC we used as a backup DC and WSUS server.
Instead of bluescreening, the system would spontaneously reboot at shorter and shorter intervals. Yes, I did all the usual troubleshooting steps, check for virii, power problems (it shares a UPS with the main server, which has worked flawlessly), unseated cards, DIMMS, etc.
The heatsink to the left is the CPU (Dell servers use ducted blowers rather than CPU fans.) There’s a noticeable convex bulge on this cap, though not as obvious as on the MSI board I showed earlier.
Both the Dell board and my MSI board were manufactured around the same time, mid-2001 or so, and that was a time when bad caps were first appearing in equipment,
The only apps we had on the box were Windows Software Update Services, and Harmini, our music library (which is just a USB harddrive loaded with production music for which we pay a license.) Both were moved to another server without any downtime.
I’m not optimistic about getting this one fixed; Dell doesn’t guarantee parts after 5 years, and with its (relatively) expensive SCSI RAID subsystem, it wasn’t worth upgrading when we had the chance several years ago. I hate throwing out servers after 5 years, but with the move to 64-bit servers, that’ll be a consideration,too.