Re: Blue screen of death when closing connections



Matthew Chapman wrote
(in article
<slrndjt98h.kcs.matthewc@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>):


> 0xab just indicates a certain type of memory leak, there are a
> number of different issues that have been reported here that are
> resulting in 0xab, mostly related to fonts. Your numbers don't
> follow any of the patterns I recognise, though.

Again, I'd be much more interested in solving the other problem
which leads up to this, but apparently that doesn't ring any
bells here.

> It would be worth trying 901150 to see if that resolves the BSOD
> (you can find a copy here: http://arizstuart.tripod.com/).

I'm not sure I want to install an unreleased hotfix on a
production system, especially one obtained from a third party
website instead of Microsoft. I think I'm going to try and
avoid it until a production update is available.

> If you want to try and diagnose it further you can do the
> following:
>
> (1) Install Debugging Tools for Windows from
> http://www.microsoft.com/whdc/devtools/debugging/default.mspx
> (2) File -> Open Crash Dump -> C:\WINDOWS\MEMORY.DMP
> (3) Type the following in the command window:
> !analyze -v
> !poolused 8
> and report the results. If it complains about symbols you may
> need to also do:
> .symfix
> .reload

I'll look into this, thanks.

> Windows can't know that this problem exists until it actually
> terminates the session, and finds that there's unfreed memory
> left over. Arguably, it should just emit a warning somewhere and
> discard the remaining memory.

Yes. Blue screening for a memory leak doesn't make sense,
unless you've leaked it all and can't recover.

> However, there might also be wider
> integrity problems at the root of the leak, which is I guess why
> Microsoft chose the "safer" route of stopping the system.

Safer for MS, or safer for my users? When dozens of people have
work in flight roll over, over 1300 bytes, that doesn't make any
sense.

>> Also, I am curious if anyone has a cure for the stuck session
>> problem which precedes this problem?
>
> I think it may be unrelated.

As I said before, it NEVER happens unless I manually disconnect
a session through the manager. It happens almost exactly timed
to the mouse click. No exceptions (no pun intended).

> Do you have any extra software
> such as antivirus etc which runs in each session?

Most are running only Office and IE currently, typically word,
excel, outlook and IE are the only apps being used.

--
_ __ _
| | ___ / _| |_ _ _
| | / _ \ |_| __| | | |
| |__| __/ _| |_| |_| |
|_____\___|_| \__|\__, |
|___/
All of God's creatures have a place..........
..........right next to the potatoes and gravy.

.



Relevant Pages

  • Re: Memory leak on ASP.NET web site
    ... about 25MB per session between SQL Server and IIS combined that is never ... regained after the Session times out. ... Anyone know what would cause such a drastic memory leak? ... no change is made to the memory and IIS and SQL ...
    (microsoft.public.dotnet.framework.aspnet)
  • Re: Memory leak on ASP.NET web site
    ... about 25MB per session between SQL Server and IIS combined that is never ... regained after the Session times out. ... Anyone know what would cause such a drastic memory leak? ... no change is made to the memory and IIS and SQL ...
    (microsoft.public.dotnet.framework.adonet)
  • ASP.NET - Detecting memory leaks
    ... How would I go about detecting a memory leak? ... our site set up to only have a root "bin" folder for .NET apps, ... The reason I ask is because occasionally, our site will start having session ...
    (microsoft.public.dotnet.framework.aspnet)
  • Re: ASP sessions expiring prematurely
    ... Make sure not to put any objects in Session or Application, ... "Larry Bud" wrote in message ... > mail@jazzis.com wrote in message ... > Not releasing objects properly causing a big memory leak? ...
    (microsoft.public.inetserver.asp.general)
  • Re: MEX calls and (continuous?) memory leaks!
    ... Thank you all for finding and reporting the memory leak in GEOTIFFINFO. ... I'm happy to report that the bug is now fixed and the fix will be available in the next release. ...
    (comp.soft-sys.matlab)