BSOD - frequent, started just few days ago

[batwing]

Hello,

This PC is about 1,5 yr old and works only as HTPC. Its main duty is Media. Movies, music using mostly Win media center and barely surf the internet to access Hulu, Netflix and similar media locations.

Win 7 64 Ultimate

I am sorry but I barely remember the specs of this PC, is a very low end machine. CPU is an AMD dual core.

It is working flawlessy for its duty and this BSD issue started just 3 days ago after a Windows update procedure which didn t work too well.
There were something like 20 updates but at the end Windows update reported 2 updates were not installed due to "unexpected error".

I attach the requested files, i hope you can see some reason for this weird behavior.

Thank you very much!

EDIT: I noticed somewhere the reporting is not finding the Antivir, however I am using Avast FREE latest version, is active and working

 

[Vir Gnarus]

Are you sure it was just only 3 days ago? While majority of your crashes do appear from just yesterday, I do see one from March 21, another from May 5, and two from May 6. I think the Windows Update may be the cause for most of the crashes you're experiencing now, but the failure to update properly may have occurred by a much deeper problem.

I noticed particularly on a crashdump from May 6 that a CPU register had a memory address stored in it that got corrupted by one solitary bit. This is very often a hardware issue, and rarely does a software bug cause a solitary bit from some arbitrary cpu register to get scuffed up.

I recommend approaching this with some hardware tests. Here's my standard fare of them:


RAM: Memtest86+ - 7+ passes
CPU: Prime95 - Torture Test; Large FFTs; overnight (9+ hours)
GPU: MemtestCL - Run twice (if any of the tests work on your GPU; ATI cards will need to install the ATI APP SDK as it requires OpenCL)
Drives: Seatools - All basic tests aside from the Fix all or the advanced ones.

Preferably you run these in the order presented. All of these (excluding MemtestCL) are included in the UBCD if you prefer a Live CD environment (which is the best environment to test hardware on). Note that Prime95 currently does not work on the newest version of UBCD. Also, please provide us temps/voltages using HWInfo with Sensors only option checked. Log two 30-minute instances: one for idle, and one for high load.

Analysts:

From 050612-14196-01.dmp:

CONTEXT:  fffff880037aa640 -- (.cxr 0xfffff880037aa640)
rax=ffff[COLOR=Red]7[/COLOR]8a002f53010 rbx=fffffa80015357e0 rcx=fffff8a00c3c8220
rdx=0000000000000002 rsi=0000000000000041 rdi=0000000000000001
rip=fffff80002fbe617 rsp=fffff880037ab020 rbp=0000000000001000
 r8=0000000000000001  r9=fffffa80015357e0 r10=fffffa8001535288
r11=0000000000000002 r12=fffffa8001535280 r13=0000000000000000
r14=fffffa800163a060 r15=000000006e664d46
iopl=0         nv up ei pl zr na po nc
cs=0010  ss=0018  ds=002b  es=002b  fs=0053  gs=002b             efl=00010246
nt!ExAllocatePoolWithTag+0x537:
fffff800`02fbe617 48895808        mov     qword ptr [rax+8],rbx ds:002b:ffff78a0`02f53018=????????????????

 

[batwing]

Thank you very much for your support Vir.

It is possible the issue happened earlier than 2 or 3 days but maybe it was not so consistent and i didn t really noticed.

I will start the troubleshooting with the tools you provided and then i will report for further analysis.

Thank you again for your time.

 

[batwing]

@Vir:

I was able to test a couple things:

Prime95, reported no issues and the temp was between 57 - 60 C
Memtest86+, reported no issues after 8 passes

Didn t have time for the other 2 tests yet.

 

[batwing]

However i still have this issue and is quite frequent.. it seems happening at least 5 or 6 times a day.

This is a media server, is always on. it seems however it happens on idle. So far never happened when i was 'using" the PC.

I have been watching movies for hours (I recently had a marathon of our fav show with my g/f) and never failed.

Instead, when i am not using it, when i get back there to use it, i find the message "Windows recovered from an uncorrect shutdown" or something like that.

I noticed from the Event Viewer, yesterday it happened at 3.37 am and at 5.59am, obviously there was nothing going on at that time, I was not using it.

I am wondering if some automatic update recurring at weird time can do this. There were no tasks of my knowledge runnin at that time. How can i see what was running at that time on my pc ??

Is there any way to make a query on Event viewer by Time to see what was running?

This thing is driving me nuts

 

[batwing]

A little update:

I am digging into the event viewer to find clues.

I found that for every crash BSOD there is a refernce to a Kernel-event Tracing reporting as following:

EventDataSessionNameMicrosoft-Windows-SetupFileNameC:\Windows\Panther\setup.etlErrorCode3221225485LoggingMode5

Now, the interesting thing is, looking at time log, there is a consistent difference of 3 seconds between crash and this log. I will explain:

BSOD at 9.45.16 AM
Krn-trace (above) at 9.45.19 am

this is recurrent for all of the instances. So my question is: is this a cause or a result? i dont get how it can be a cause if it happens 3 seconds after crash.
However, 3 seconds after crash, there is not even time to PC to reinitialize a POST and get back online (I wish ), so, have to be something weird in the way it get logged.

I am focusing on this cause because I am having weird error messages while trying to updated 3 specific updates ( i have to find the 3 updates IDs, not available right now) and this Log is directly talking about Windows Setup which is what is failing during the updates.

However, I tried manual updates to force a crash and nothing happen. I just get an error message from Win update window and that s it, no crash.

What is that Panther\setup.etl ?

I recall I found something googling with user saying is possible delete that folder, or file and it will be recreated.

Waiting for suggestions, please

 

[Vir Gnarus]

That does seem rather odd. As for Panther\setup.etl, it's the Windows Update performance log, which should harbor a good bit of details on what was going on at the time, as mentioned here. If you can, zip it up and send it over. Also while you're at it, can you please take a screenshot of the error message(s) you're receiving? There may even be logs associated with them as mentioned in the article I linked too. See what you can scrounge up. The more information the better.

Anyways, I still don't understand how it has anything to do with the single register bit corruption I witnessed prior. I recommend continuing the rest of the hardware tests, as well as doing Prime95 again on Blend this time instead of Large FFTs (again, same length of time). Before you start any of them though, I recommend you provide the HWInfo logs as previously requested. Those can assist us on determining if this is a PSU problem. However it's not definitive: there are no definitive hardware tests for the motherboard and PSU aside from a multimeter and a mobo diagnostic kit. So if it's either of them causing trouble, either most or all hardware tests will show up probs, or no hardware tests will show up probs.

In addition to all this, to make sure we aren't dealing with some driver-based problems, turn on Driver Verifier. Any crashes you experience while it's on will give us better crashdumps to look at. Do not select the IRP Logging, Force Pending I/O Requests, and Low Resource Sim checks. Follow all other instructions in totality. Provide us any new crashdumps.

 

[batwing]

Hi Vir,

I do thank you again for your time on this matter. i do apologize it took a while to reply. I was out of town for a project.

Finally I am back and and I am back troubleshooting this weird issue. Some updates:

I am having constant crashes, on an average of at least 5 random crashes perday. Is not something I can repeat. I had crashes also while I was using the machne, watching a movie Usually was happening when the machine wa in idle).

I re-ran Mem86 full night for a total of 9 hrs, no error whatsoever. I tried to run Mem86 "probing" scan and it crashes immediately. I have no idea what is the "probing" option tho..

I am re-running Prime95 in "Blend" while typing, I will report results.

I was reading all of the "Bugcheck" repoprts in Event Viewer and although they are usually reporting some different code, reading the code explanation, it result always being a code related to some memory fault.

I found a few instances of bugcheck code 050 and searching the internet it seems this may be related to some driver failure.

When Prime95 will be over i will try to activate the Driver Verifier. I was also thinking if as a final resource it would worth a shot just "repairing" Win 7 with a full install on itself.

I did that in the past when I was having weird issues following these procedures:
http://www.sevenforums.com/tutorials/3413-repair-install.html

If some sorta of drivers are failing (plus that weird update behavior I am having) maybe a nice "cleanup" would help better instead of having an headache troubleshooting, thought?

 

[Vir Gnarus]

I believe the probe method is to determine RAM size. It's antiquated and is available just for use on really old RAM types.

So far it sounds like it's ending up to be the motherboard or PSU. Often in real-world scenarios they'll bug things out but never do doing specific hardware tests. A single bit corruption is a common symptom from them. Again, you'll want to finish all the aforementioned stuff like providing that HWInfo log and turning on Driver Verifier.

As for figuring out the bugchecks, understand that it takes a good bit of knowledge to properly diagnose these kinds of things, and even then you can only go so far to determine if it's software-derived (in which case you can find exact cause) or hardware-derived (where you can only make an educated guess what hardware it is). 0x50 bugchecks are one of the most common bugchecks of all and can come from nearly anything so unless you whip out a kernel debugger and start cracking on a kernel or full dump or two, you most likely won't be able to determine cause.

Driver Verifier can at least greatly assist in determining if this is software-derived or otherwise. Provide us any new crashdumps after you turn it on. Remember which checks not to select.

 

[batwing]

Thank you Vir,

I decided to go for the full clean install, to speed up the process. Bottom line, if it fails again on a total fresh install, it has to be Hardware.

In fact right after the fresh installl, I got the same random crashes, with still similars bug checks.

At this point, although Mem86 didn t report failure, considering Prime on "Blend" fails with a crash after an average of 20 minutes (i tried twice and i found the machine crashed and event viewer reported the same issue and it happened about 20 minutes since when I started the Blend test), I will start removing the RAM and using 1 deck at the time and see if one of the 2 works.

Funny to say, I can t event recall right now if I have 2GB RAM or 4GB RAM on that machine, therefore I am not even sure if I have 2 or just one deck of RAM.

Well, i guess I will discover pretty soon.. and if i have 1 2GB RAM stick, then I ll buy another one and I will test when they will deliver

Thank you so far, i ll keep you posted... I am scared to think about the MB issue, mostly because is just very annoying having to open the chassis and remove everything

 

[batwing]

Update:

first of all I have 1 stick RAM only: 2GB ddr2 800

my MB is:Newegg.com - ECS A780GM-M3 AM2+/AM3 AMD 780G Micro ATX AMD Motherboard

Its manual says I should be able to switch the RAM stik in either one of the 2 available slots, however it doesn t work like that.

When I unplug the stick from the actal slot and i place it into the otherone, the MB will not even POST.
If I just remove the RAM completely, MB will POST and will report beep codes for missing RAM.

So I did put the stick back and the original slot and the MB will POST normally. Almost normally, because however at the first 2 attempts to POST and get into OS, the MB was stuck at the very first part of the video info on screen, unable to even get into BIOS settings.

After 2 attempts MB started POST correctly and was able to get into BIOS correctly. The issue didn t repeat anymore. After that always loaded Windows correctly. Howevercrashes now seems to be way more frequent.

I already ordered another stick or RAM. My path at the moment is testing a new stick of RAM first (cheap enough) and then if still is having issues I will replace the MB. I start to believe is the MB, however an extra stick of RAM is never a good idea to have

I actually do not believe there is any PSU implication in all of this, mostly because as i said, manual says I could set the ram in either one of the 2 slots, but is not happening, leaving me to think the MB controller is at fault.

 

[writhziden]

Sounds like either the Motherboard, CPU, or RAM. Your troubleshooting steps are sound, so see if the new RAM resolves it. If not, make sure all your power connections are set up correctly to the motherboard including the fans. I had a NB fan and CPU fan power junction swapped once and it caused some very strange behavior and my system would not get past POST.