Question Locked

W2k3 SP2 reboots when a remote connection logs off

0 helpful
10 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

Hey guys

Having an issue with a windows server 2003 build.

It's been stable and working fine for the past year. However since last week, everytime a user logs off their terminal session, the server immediately restarts (just as if somebody pressed the reboot button on the tower - no windows shutdown dialog).

To my knowledge, no new programs have been installed other than the normal windows updates since this issue started.

There is a user logged in to the local console at all times as well and there is no specific event in the event log other than event id 6008 after the fact and there is no time for a memory dump since it is an instataneous shutdown.

Any suggestions would be appreciated.

Thanks.

PS : I have yet to find out how to give out the "helpful answer" rating on this site. If you can tell me how, I'd be hapopy to flag helpful responses as such.

10 replies

by shasca

Check to see under System applet.

Disable automatic restart on system failure.

This will at least give you a chance to find what is causing the problem.

by rob miners

Minidump Files can be found here. C:\WINDOWS\Minidump\Mini******-**.dmp

How to read the small memory dump files that Windows creates for debugging

http://support.microsoft.com/kb/315263

To download and install the Windows debugging tools, visit the following Microsoft Web site:

http://www.microsoft.com/whdc/devtools/debugging/default.mspx

Instructions on using Windbg.

Open Windbg and select file, Symbol file path and brows to the Symbol folder that you have downloaded and installed Symbols to, select OK.

Close the workpage and save the Workspace information. This should lock in the Symbol path.

Open Windbg and select file and select Open Crash Dump then navigate to the minidump, highlight it and select Open.

There are two ways to use !analyze -v the easiest is to click on !analyze -v under Bugcheck Analysis.

When you have ran the initial dump if you look to the bottom of the screen you will see kd> to the right of that type in !analyze -v and press the Enter key.

Ctrl + a will let you copy the information and paste it into notepad.

Look to the bottom of the page for information like this. This was fixed by updating the Graphics Drivers.

FAULTING_THREAD: 864f6a98

DEFAULT_BUCKET_ID: GRAPHICS_DRIVER_FAULT

CUSTOMER_CRASH_COUNT: 1

BUGCHECK_STR: 0xEA

PROCESS_NAME: logon.scr

LAST_CONTROL_TRANSFER: from 00000000 to bfef0fea

STACK_TEXT: f6434e6c 00000000 00000000 00000000 00000000 sgiul50!DrvBitBlt+0x585

STACK_COMMAND: .thread 0xffffffff864f6a98 ; kb

FOLLOWUP_IP: sgiul50!DrvBitBlt+585 bfef0fea 75f9 jne sgiul50!DrvBitBlt+0x580 (bfef0fe5)

SYMBOL_STACK_INDEX: 0

SYMBOL_NAME: sgiul50!DrvBitBlt+585

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: sgiul50

IMAGE_NAME: sgiul50.dll

DEBUG_FLR_IMAGE_TIMESTAMP: 3b7dfec7

FAILURE_BUCKET_ID: 0xEA_IMAGE_sgiul50.dll_DATE_2001_08_18

BUCKET_ID: 0xEA_IMAGE_sgiul50.dll_DATE_2001_08_18

When you open the post from the contributor you will see Mark Helpful near the bottom.

Keep us informed as to your progress if you require further assistance.

If you think that any of the posts that have been made by all TR Members, have solved or contributed to solving the problem, please Mark them as Helpful so that others may benefit from the outcome. ;) :D

by optronix_quebec

Thanks for the info.

It wasn't dumping the memory because I had moved the paging file to another drive than the default windows one. As such, it refused to even show the stop error and to debug. I've created a small 16 MB one on that drive now so that portion is fixed now.

I also ran through the debug file info but as I've never used that tool before it's all a bit confusing. I've posted the log below.

From what I can gather it would seem to be caused by a missing (or corrupted) file relating to crsss.exe. I've also got a ghost of the server that is about a month old in an offsite location. No major changes have been made to the system other than the windows updates once a week so I could most likely use that version for replacement files if needed as I'd rather not use windows repair. It gave me spotty results at best in the past.

Any suggestions ? From my limited understanding, it does not seem to be driver related.

-----------------------
SESSION_HAS_VALID_VIEWS_ON_EXIT (ba)
Caused by a session driver not unmapping its mapped views prior to a
session unload. This indicates a bug in win32k.sys, atmfd.dll,
rdpdd.dll or a video driver.
Arguments:
Arg1: 00000001, session ID
Arg2: 00000001, number of mapped views that are leaking
Arg3: e7350470, address of this session's mapped views table
Arg4: 0000001f, size of this session's mapped views table.
Debugging Details:
CUSTOMER_CRASH_COUNT: 1
DEFAULT_BUCKET_ID: DRIVER_FAULT_SERVER_MINIDUMP
BUGCHECK_STR: 0xBA
PROCESS_NAME: csrss.exe
CURRENT_IRQL: 0
LOCK_ADDRESS: 808a79a0 -- (!locks 808a79a0)
Resource @ nt!PiEngineLock (0x808a79a0) Available
WARNING: SystemResourcesList->Flink chain invalid. Resource may be corrupted, or already deleted.
WARNING: SystemResourcesList->Blink chain invalid. Resource may be corrupted, or already deleted.
1 total locks
PNP_TRIAGE:
Lock address : 0x808a79a0
Thread Count : 0
Thread address: 0x00000000
Thread wait : 0x0
LAST_CONTROL_TRANSFER: from 8092caae to 80827c83
STACK_TEXT:
ae588c4c 8092caae 000000ba 00000001 00000001 nt!KiServiceTable+0x1c7
ae588ce8 8084c795 884aa838 00000000 88723db0 nt!NtQueryDefaultUILanguage+0x5
ae588d04 8094c4eb 884aa838 88723db0 88723ff0 nt!MiSetPageModified+0x1bd
ae588d8c 8094c63f 00000000 00000000 88723db0 nt!PiProcessQueryRemoveAndEject+0x1a6
ae588da4 80949b8a 88723db0 00000000 00000001 nt!PiProcessQueryRemoveAndEject+0x3ce
ae588ddc 8088e092 bf9196c2 88e6c178 00000000 nt!IopGetDeviceResourcesFromRegistry+0x205
00000000 00000000 00000000 00000000 00000000 nt!ExReleaseRundownProtectionEx+0x47
STACK_COMMAND: kb
FOLLOWUP_IP:
nt!KiServiceTable+1c7
80827c83 5d pop ebp
SYMBOL_STACK_INDEX: 0
SYMBOL_NAME: nt!KiServiceTable+1c7
FOLLOWUP_NAME: MachineOwner
MODULE_NAME: nt
IMAGE_NAME: ntoskrnl.exe
DEBUG_FLR_IMAGE_TIMESTAMP: 48a2ac75
FAILURE_BUCKET_ID: 0xBA_nt!KiServiceTable+1c7
BUCKET_ID: 0xBA_nt!KiServiceTable+1c7
Followup: MachineOwner

----------------------

PS : Thanks for the "mark as helpful" info. I've been looking for that button for ages but it's been starring me right in the face this whole time!

by rob miners

since you have updated your Device Drivers.

How to Use Driver Verifier to Troubleshoot Windows Drivers

http://support.microsoft.com/kb/244617

Keep us informed as to your progress if you require further assistance.

If you think that any of the posts that have been made by all TR Members, have solved or contributed to solving the problem, please Mark them as Helpful so that others may benefit from the outcome. ;) :D

by optronix_quebec

It's been a while but I usually abide by the saying "if it ain't broke, don't fix it" in regards to drivers for the server so I didn't really see a reason to.

I'll however take these further steps before reporting back :

- update to the newest drivers, if any
- run verifier.exe to verify all drivers are functioning as they should
- run memtest to rule out RAM issues
- test my HDD with the WD Diagnostics utilities as this is a raptor drive

From what you guys seem to be hinting toward, this would seem to be a driver/hardware issue based on the posted log so I'll check that out and that should hopefully fix the issue.

Thanks for the help so far.

by rob miners

the case these days, you have to keep on top of device drivers. Keep everything current and up to date.

by optronix_quebec

I've went through the paces and the HDD and Ram were fine.

I did find driver updates to a few components but when I ran verifier it still had issues as it did a BSOD at startup. The BSOD didn't point me toward any particular driver or dll. It was just stating a "faulty driver on kernel stack". When I went to check verifier's settings, it didn't point me toward any particular faulty device either.

I however think I've figured it out.

One of the printers on one of the workstations that connect in terminal services to the server has been behaving oddly as of late (pages getting stuck in the spooler, never printing, etc.). Since the crashes occured at logoff when TS "unplugs" its printers on workstations, I thought it might have been the printer crashing the system.

I've deleted the printer from the computer and shut it down and the system hasn't crashed since.

It's only been a few hours but I had people test 4-5 logon/logoff cycles per workstation to see if a crash would occur and it's been solid so it would seem the HP printer was the culprit.

I'll post back here in case the issue ends up being different.

Thanks all for your help.

by thumbsup2

... never install HP drivers offered through MS Update. Each and every time I've tried it, either the printer quits functioning or the system crashes when you try to use the printer.

So, while you're looking at this particular printer issue, be sure to get the latest drivers directly from HP and ignore any drivers offered by MS Update.

This discussion is closed to new replies.