DCOM CRASHING MY COMPUTER!!!
Mark
someone can help, so lots of information.
Runnig a Dual Opteron DualCore workstation , 8G of RAM,but I have spent a
couple of weeks trying to stop x64 from crashing, but now I have done a fresh
install and apart from hardware drivers the only other program is a adware
scan.( I am operating behind a router NAT and a service provider NAT)
The following is what I can find out after a re-boot, which sometimes is
done by the computer or by me if it locks up.
EVENTVIEWER:
SYSTEM:
0.41seconds: Microsoft(R)Windows(R)5.02 3790 Service Pack 1 Multiprocessor
Free.
0.41seconds: Event log service was started.
0.41seconds: The COM sub system is suppressing duplicate event log entries
for a duration of 86400 seconds. The suppression timeout can be controlled by
a REG-DWORD value named SuppressDuplicateDuration under the following
registry key:P HKLIM\Software\Microsoft\Ole\Eventlog.
0.47seconds: The Application Experience Lookup service started successfully.
0.48seconds: ERROR: The machine-default permission settings do not grant
Local Activation permission for Com Server application with CLSID
{555F3418-D99E-4E51-800A-6E89CFD8B1D7} to the NT AUTHORITY\LOCAL SERVICE SID
(S-1-5-19). This security permission can be modified using the Component
Services administrative tool.
0.48seconds: ERROR: (The above is repeated)!
NOTE FROM ME. This refers to : REG Z C\Windows\System
32\rundll32.exe.....Run DLL as an application. ( Within the Windows file it
is the only dll file that does not have an dll icon , it's icon is a blank
page)
The timings jump back at this point:
0.24seconds: The IPSec Driver is starting in Bypass mode. No IPSec security
is being applied while this computer starts up. IPSec poloices, if they have
been assigned, will be applied to this computer after the IPSec services
start.
From this point the computer goes into start up.
In APPLICATION VIEWER:
0.47seconds: The EventSystem sub system is suppressing duplicate event log
entries for a duration of 86400 seconds. The suppression timeout can be
controlled by a REG_DWORD value named SuppressDuplicateDuration under the
following registry key: HKLM\Software\Microsoft|EventSystem\EventLog.
0.50seconds: The description for Event ID (1800) in Source (SecurityCenter)
cannot be found. The local computer may not have the necessary registry
information or message DLL files to display messages from a remote computer.
You may be able to use the /AUXSOURCE = flag to retrieve this description;
see Help and Support for details. The following information is part of the
event.
I do not know what information it refers to as it ends there.
0.49seconds: Performance counters for WmiApRpl(WimApRpl) service were
removed sucessfully. The Record Data contains the new values of the system
Last Counter and last Help registry entries.
0.49seconds: Performance counters for the WmiApRpl(WmiApRpl) service were
loaded sucessfully. The Record Data contains the new index values assigned to
this service.
SECURITY VIEWER:
Success Audit (through out the time frame......no errors)
That is it, the only other problem is when I open the windows Security
Center it cause's broken lines across both screens.
If anyone can understand this PLEASE let me know.
Mark
Tony Sperling
I really shouldn't be answering this, because I do not know. Something in
there mentioned 'permission settings' - are you shure you're alright there?
But Event ID's can be searched for in the Knowledge Base, and so can
keywords in the error messages themselves.
Did that machine ever run anything else, besides win64, or is it deployed
specifically for this purpose? Could it be in need of drivers or BIOS
updates for you specific networking Hardware?
Happy hunting!
Tony. . .
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:2B9959EC-568D-473C...@microsoft.com...
Mercury
The eventlog entries are somewhat normal or IMO no indication of an
association with a crash. Min you that depends on the definition of
"crash" - see below. Perhaps there are two problems?
What stop code(s) do you get? Do you get PM dump? There hould be an entry
in the eventlog for the stop code.
If the machine hard locks - mousepointer won't move, caps lock key doesn't
work and all activity has ceased, then I would say it is hardware.
A component could be 'racing' from error to error and cause a drastic
slowdown. On a multi CPU system I would expect the system to still be
responsive unless it locked some exclusive system resource.
You can check for the COM component via Admin Tools, component services and
look for 555F3418-D99E-4E51-800A-6E89CFD8B1D7 there or search the registry
for the component and find out what it is associated with. Ok is that what
you mean by "This refers to : REG Z
C\Windows\System32\rundll32.exe.....Run DLL as an application. " What is
rundll32 trying to run?
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:2B9959EC-568D-473C...@microsoft.com...
Mark
And yes Mecury you are right about the hardware issue. I did not explain
myself correctly in that I knew DCOM was not at fault, DCOM Server Process
Launcher in Services has a defalut setting to shut down the computer on First
Failure , it was a question of what was causing the failure. As for the
Rundll my back tracking was at fault as the entry actually refers to a
registry entry which then leads to its proper DCOM Config association :
Registry:
Local Service: ab(Default) REG-Z ShellHWDectection
abThreading Model REG-Z Free
Prg ID: ab(Default) REG-Z HWDeviceCustom Properties1.
Version Independant Prog ID. ab(Default) Reg-Z HWDeviceCustom Properties.
DCOM (association) HWDeviceCustomProperties.
How do I work out which piece of hardware it is refering to? I do not see
any stop codes or PM dumps, unless it is in plain English I may miss
something, everthing I put in my first post was how it was in the event logs
around the time of the problem I did not leave anything out .
In my first post I mention the " Suppression of Duplicate Event Log
entries", in DCOM Config there is Sens Logon Events and Sens Logon Events2
both with the same Application ID, is this normal?
Just to let you know, this is a new workstation I have built for x64 and
nothing else, all the drivers are up to date, I use a wireless bridge to the
home router, where all the home computers access the net but not each other(
there is no home network set up in the true sense of a network).
Regards.
Mark
Mercury
Try directx video tests - that will bring out any graphics related issues. -
dxdiag.exe.
Any other h/w speicifc errors by trying to provoke them eg network copy some
large files repeatedly, many many small files repeatedly...
check copu temps and psu voltages as you go.
Then prime95 for a final stability check.
sorry, been off air.
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:157799F9-AA18-4F3C...@microsoft.com...
Mark
Where do I find Prime95, the web site I found was a site looking for new
prime numbers, but I did not see a download for memory testing specificaly,
just downloads to become part of the network search for new primes.
I have tested for virus's as I was intermitently finding CLSID followed by
a few numbers (not a full string I could make sense of)coming up in black
text at the top left hand corner of my desktop. Very strange!
Will try what you have suggested. Thank you very much for getting back to
me, I appreciate the help.
Mark.
Mercury
prime95 is at www.mersenne.org you will want ot run the torture test.
try anti spyware procedures as well.
you said dual core, dual cpu opteron? This implies ECC / or registered
memory, so check ram ratings and bios settings - I would expect this to be
really stable.
You could check for a motherboard specifc news group, or if you can't find
one try posting at alt.comp.mainboard.periphs.asus - but say straight up if
you don't have an Asus mobo - they are a good bunch there and often help
'others'.
But lets not get off the beaten track. You could try a progrssive strip down
or build up to see if a h/w item is causing the issue. Check BIOS version,
run sigverif.exe to check all drivers and check for updates, also try to
nail down the circumstances under which the crash occurs.
- Tim
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:F6AD424C-65C4-496A...@microsoft.com...
Wayne Wastier
Mark
Mark
Mark
Just to let you know, prime 95 or memtest 86+ do not work as memtest does
not support multi cpus or multi dual core cpus and Prime does not advise
using itself as a true memory tester as it cannot recognise a lot of memory
faults. Memtest just hangs on test 4, normally crashing of memtest
represents a faulty cpu but as it does not support dual cores, I am not sure
this is the case.
As far as drivers, bios etc are concerned I am up to date, there is a
revised version of my bios but apparently it has it's own problems causing or
falsely reporting cpu tempreture rises by 20c. Any improvements it offers, do
not relate to the problems I am having.
I noticed during a fresh x64 installation the screen would flash from the
installation screen to black during device and hardware installation( is this
normal), some programs I have installed also seem to have not been installed
fully, with windows updates not showing at all in add remove programs when
'show updates' is chosen.
Many thanks for all your advice, I think it is time to push components back
to the manufacturer's, as I have been fighting this battle for six weeks now.
I paid a lot of money in the hope that at least everything would work. Well
that was flawed logic.
Gotta laugh, but maybe not just today.
Mark
Tony Sperling
multi-cpu's or whatever. Might be worth a shot?
http://oca.microsoft.com/en/windiag.asp
Regards, Tony. . .
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:75655B83-0D8B-4199...@microsoft.com...
Tony Sperling
4gig's of memory, but swapping the sticks in their sockets and running it
twice should give you an indication.
I also became aware of the rest of your post - the flashing screen, I would
say is normal, and some machines can show a more violent flashing than
usual. About software installations, I forget, but I think MS issued an
updated installer a while ago, if you are on Windows Update, that should
come automatically. AND, 16bit installers do not work at all.
Tony. . .
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:75655B83-0D8B-4199...@microsoft.com...
Mark
I will give the Microsoft test a try but I think I read on another support
site that it does not support Dual Core's. I am talking to a company called
Micro 2000, they have a specailist diagnostic tool and they are looking at my
specs to see if their
tool can do the job properly.
Reluctant to start pulling my DIMM's in and out due to their cost, if the
Microsoft test works, then maybe it will be worth it.
Mark
Mark
I have ran the Microsoft Memory test and it has come out fine for the first
4GB, any idea which memory it has tested. Would that be the first 4GB for CPU
1 or the first 2GB for CPU 1 and CPU 2. Sorry ,but I cannot tell by the test
screen and it does not say anyhthing on the site.
Secondly I have ECC REG memory, the test will only spot any errors that the
ECC has not been able to correct. Would it be an idea to run the extended
test?
Mark.
Tony Sperling
( I don't have any such thing as yet! ).
I would believe that it don't much care about the processors, it tests the
memory from the begining to it's limit. It would see that as a pool - your
question really is "how will the processors divide the memory-pool between
them?"
I think you should start a new thread, with that as the subject, since the
DCOM keyword is rapidly becoming obsolete.
Tony. . .
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:FAACF3BF-F1BE-4C62...@microsoft.com...
Mercury
remining CPU to test the ram?
The objective of Prime95 torture test is not to test ram, but to test
overall stability IE CPU, Cache, RAM, intervening chips (few in your case).
A description I read was along the lines of running the algorithmns in an
intense manner knowing what the outcome will be -it checks the outcome and
if it is wrong ==> something is wrong. IE it is often used as the last
system stability test when building.
I'd do a strip down to minimum - one CPU (then the other later), test 4gb
ram at a time (pull out the rest), keep an eye on thermals and PSU the whole
time. Remove all extraneous peripherals, PCI & PCIe -with only 4gb instlled
and 1 cpu test each 4gb chunk of ram. Swap CPU chips, do the same tests
(this will take a while) - you don't need to test RAM for each CPU although
since the CPU has th interface built in some testing may be worth while.
If there were an odd CPU0 failure, swapping CPU's sockets may allow the OS
to run and report the failure.
If you can test each device that plugs in on another system, do so, and put
them aside until the fault is rectified. Disconnect all optional internal
connectors EG front panel USB 2 and firewire, make sure no USB Disc /Key or
other devices are connected. In short the best strip don i the crdboard
test - remove the motherboard and place it on piece of cardbord. Plug in
only what is needed to get it running as above and test / build up from
there. (Don't use anti static mat as it is condutive).
Commercial system builders should have RAM testing devices - perhaps get
them to test it, but I would expect the OS to report ECC single bit
failures - not DOS though :)
You may be able to disable ECC in the BIOS. This may enable a "real" ram
test...
Can you try another PSU? I would put that high up the checklist. PSU's cause
faults like this. What rating is the PSU - amperage for each voltage & other
dpendancies. Are you exceeding any PSU abilities? Draw up a table.. If you
have a Digital Volt Meter, use it on a spare connector to verify run time
voltages. Check the reset switch works - they often stick or can cause all
sorts of issues..
The thing I failed to say clearly the other day is "what does the evidence
to date point to?" and "what evidence / concrete symptoms do you have so
far?", "What facts do you have?". I just fear that we may be leading you
down a totally irrelevant path.
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:31C5CFE1-620E-4A62...@microsoft.com...
Tony Sperling
Tony. . .
"Mercury" <m...@spam.com> wrote in message
news:dj922o$vef$1...@lust.ihug.co.nz...
Mark
I looked at the the DCOM issue and a lot of the error messages are normal
and to do with networking, not something that should be crashing the system,
work arounds also exist for these issues.
As far as the PSU (650w) is concerned, it maybe faulty but I am not over
loading it, or asking it to do anything it cannot do. But I understand that
if nothing else is found to be the cause it needs to be looked at very
closely.
REAL EVIDENCE?
Pointing in any direction, graphics. Random lock-ups and re-boots do happen,
but I have to be using the computer, left idle and with a small amount of use
it will stay on for days and not suffer any problems.
BUT during this period there is one thing I can always provoke: If I go into
dual screen mode and expand my desk top to the other monitor and open up
Windows Security panel, white lines will always break across the center of
both screens. Expanding a window to full screen and back again will remove
the lines in what ever screen the window was expanded in. This happens
without fail when in dual screen mode.
In single screen mode Avast Anti-virus UI will suffer a graphical error and
lock up or re-boot the system, during a virus scan. But a complete disk
check (fix all errors and recover bad sectors) sorts this out.
Re-boots and lock-ups start to happen as I load more utilities and
applications and start to use the computer. I have blamed several utilities
and applications but if I do not load them or uninstall them it will happen
sooner or later with some other program.
CLSID string appearing in the top left corner of the desk top in black text.
This problem seems to attach itself to a program and starts to happen after I
start using and loading more programs. Firstly it was Quicktime Player when I
opened it, so it was uninstalled, now it happens when TweakNow RegCleaner is
opened.
Starting a program or putting in a certain disc may also provoke a re-boot
or lock-up, but this is random and does not happen everytime I use the same
disc or program. The rest of the time I could be doing anything and
experience a lock-up or re-boot.
Lastly TweakNow's PowerPack which is enabled for XP PRO x64 will install but
never load, 'Customize Exe. cannot be loaded as it is incompatible.'
Thats everthing.
Mark
Tony Sperling
increasingly more about virus infection, can you set up avast to
scan-on-boot before Windows loads, and without it breaking?
You probably already did, I know, but the PSU being 650 W only tells us how
much it can deliver, not overloading it doesn't have to mean that it is
doing it's job properly. Good quality means a well enginered construction
from first rate components, and being well tested before delivery. This
often translates into 'expensive' - compared to standard computer equipment,
but not really costly compared to getting the job done. So, I would try and
get a second opinion on the quality aspect of that PSU - also, how is the
'quality' of the current supplied to your house? That is not always in
order, if you live at the outer extreme of your power stations range - or,
indeed too close, that could mean trouble as well.
Random behavior from electronics very often can be isolated as originating
in power, heat or frequency problems - you also could try and underclock the
thing slightly, if your BIOS allows that. Not that I think it is running
over it's rating, but something in there may not be happy with running at
that specific rating. Is the motherboard fastened with spacers all the way
around and not having substantial sections unsupported? If you touch it,
will it flex or is it rigidly secured?
I doubt it if looking for memory related issues will get you an answer
quickly, you might be more 'productive' by making exclusions at this stage.
It might sound drastic, but I would put that machine on a well lit table and
I would yank and re-insert every cable and connector and card (including
memory) in there one by one and I would exchange the IDE flat cables as they
are relatively cheap, during this inspection much information can surface
about the quality of components, and you will strengthen your 'relationship'
with your machine. It may not solve your problem, but you will learn from
the exercise and you are sure to benefit at some later time.
I would really hate it,if at the end of all this, it turns out you were
being led down the wrong path, as Mercury said. Frightening thought! But, at
least, you will not be bored this week-end, I suppose. Good luck, Mark.
Tony. . .
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:CF5B3C4B-6B62-4D38...@microsoft.com...
Mark
I have already changed the IDE cables and switched around the hard drives
and down fresh installations. I have used old drivers and new drivers but
still the same results.
So I finally installed XP PRO 32 just to see what would happen and I got a
stop code for the first time!
"Thread stuck in device driver"
"This issue may occur if the display driver is caught in an infinite loop
while it is waiting for the video hardware to become idle."(Could this be the
suppressing duplicate log entries, which in x64 was related to a custom
device?)
"This issue typically indicates a problem with the video hardware or that
the display driver cannot program the hardware correctly."
Is this it!!!!? I am doubting it is the driver as I have used 4 versions in
total fo 32 and 64....or is it possibly something else is just messing with
the drivers. OR ....can it be said that there is something wrong with the
graphics card.
I have not spent much time on XP PRO 32, but the lock-ups, re-boots and a
graphical error on Avast virus scan just before a lock-up have already
happened. I am not getting the Windows Security Panel causing the stripes
across dual screens when opened.
What do you think guys?
Hope you both had a good weekend!
Mark
Charlie Russel - MVP
2.) turn off any overclocking
3.) replace that video card.
4.) Update your BIOS
I don't think it could be clearer -- you have a _hardware_ problem.
--
Charlie.
http://msmvps.com/xperts64
Tony Sperling
re-seat it, checking that everything is O.K. I assume this is a modern card,
so check the power connections, while you are in there, you may need a
DIRECT connection from the PSU, besides the one on the card itself - these
'mothers' can be hungry! If nothing surfaces, try and borrow a replacement.
I go with Charlie on this, it is not the drivers.
Tony. . .
"Mark" <Ma...@discussions.microsoft.com> wrote in message
news:2D139C8B-C65F-43F1...@microsoft.com...
Tony Sperling
stumbled on his 8GB of it, the MS utility would only swallow 4GB and then
the question surfaced how to interpret the test as he is running two
processors (two sockets), have you any idea how they would divide the memory
resources between them? More to the point, is the first 4GB of memory evenly
spread between the processors, or assigned to one of them only - or is it
random, even?
Tony. . .
"Charlie Russel - MVP" <cha...@mvKILLALLSPAMMERSps.org> wrote in message
news:OrzrtwO2...@TK2MSFTNGP09.phx.gbl...
Charlie Russel - MVP
install DIMMs.
It's a hardware or BIOS problem, clearly. If the test can't see more than
4gb (and I'm not surprised), then don't put in more than 4gb and go to the
next step.
--
Charlie.
http://msmvps.com/xperts64
Mark J.
other users located at
http://www.planetamd64.com/lofiversion/index.php/t15624.html are having the
same DCOM problems. This is a matter of "which came first the chicken or the
egg?". Does the DCOM error cause the hardlock, or does the hardlock spawn the
DCOM error. No other error messages are recieved in the system logs at the
time of the lockup. I started noticing my DCOM errors after checking my
system logs after each time World of Warcraft and EVE online crashed. Every
time those games hardlock, the system log indicated a DCOM error occured at
the exact same time. I tried every big bad benchmark (Sysmark, PCmark, Sysoft
Sandra, Seti at home running on both cores), burn in, and the windows memory
tester. Not one failure or error to be found. I performed hours of continous
gruelling tests, and ensured my hardware was tested to its limits under the
hottest conditions. But my reasoning for assumming that the DCOM errors are
at fault, they happen regardless weather I am playing those games or not. It
is only when I play those games and an error occurs thats when I have the
lockup. Simple cause and effect puts the DCOM errors at fault. Now if there
is a way to stop the DCOM errors or figure out what is causing them. Then we
may be able to make rogress. It is definitely software related, because the
occurance of a broad range of hardware platforms, but one thing that is
common, Windows XP Pro X64.
Tony Sperling
while ago, since this is an orphaned 'Re:' and I don't have access to
anything that was discussed before. You are not the same Mark as the chap
allready in this thread, are you?
I can say as much, I have not resolved my own issue yet, but it is also not
any immediate problem, it's just the errors that sits there in the log.
I suggest, if you are having any trouble, that you simply start a new thread
and we'll take it from there, all right?
Regards, Tony. . .
"Mark J." <Ma...@discussions.microsoft.com> wrote in message
news:4BF7074A-8153-4300...@microsoft.com...