|
United States1080 Posts
So for the last 5-7 days I've been getting the weirdest crash symptoms I've seen in like 15 years. This is roughly what will happen:
1) A couple of programs will lock up (a game, a web browser, whatever). During this point in time, any open comm programs (skype, vent, mumble, etc) will still be functioning.
2) After a little bit the OS decides fuck it, and I get a blue screen. This blue screen lasts a really, really short amount of time (can't even grab a picture). I noticed a couple of times some text at the bottom saying something about core dump failed, but it's only up for a short period of time so I can't quite tell.
3) After the quick BSoD screen, the comp restarts.
4) At this point, bios has decided that it would like to change the boot order of my drives. Instead of booting off of my SSD, it boots off of a HDD that has nothing but pictures and an old grub install from 5 years ago (no OS).
5) I can open up bios now and change the order of drives back, and everything will be working for between 5-48 hours, until it crashes again (repeat steps).
Interestingly bios only changes the boot order - my clock doesn't get reset, OC settings & other tweaks remain untouched.
Windows/Minidump is empty (not that I was expecting it to have anything with such a quick BSoD).
Updated my bios, no change.
Tried reverting my overclock settings, no change.
Temperatures are all normal, CPU load is very low when this occurs (<10% load, <40/50C temps).
PC specs:
Crucial M4 SSD, 120-240GB, somewhere around there
WD 640GB HDD (I think one of those shitty green drives, so slow)
4x Seagate 1TB enterprise
i7-2600k, asus p7z77-v
seasonic bronze psu, somewhere around 500W don't quite remember
some amd/ati 7xxx gpu, 7950, again don't remember
Can provide more info as necessary.
|
Did you disable the page file to save space on the SSD or something? The memory dump for the BSOD fails if the page file is too small for your crash dump settings. For "mini dump" settings, you only need a 16 MB page file, so that's what you can change to if you currently have your page file disabled.
I think you still get an entry in the event viewer about the BSOD even if the dump fails, so look around in there. It might mention the code number somewhere but there's no details like in the mini dump regarding what drivers were active and whatnot.
|
United States1080 Posts
I did not disable the page file, have 100GB open space on SSD.
Event log from last crash (5:35:51 am according to subsequent system logs) shows nothing, empty for that entire time range  Closest event is an hour away, but it's just an uninteresting adobe flash updater turning itself on & off event (occurs every hour).
|
Run BlueScreenView and see exactly what bsods you're dealing with. We can go from there when you have this.
The other thing is combing thru Event Viewer and looking for events at the same timestamps as the crashes.
|
United States1080 Posts
Will try BlueScreenView - already combed through event viewer, nothing there.
Hang on - reading the docs for BlueScreenView, does it work if there's no information in windows\minidump?
It appears as though BlueScreenView won't help, because, as mentioned in OP, there are no minidump files.
|
Lol just run it - would have taken you less time to do that than debating on whether to run it at all :p
Also sounds like you still have it set to Automatically Restart upon failure - I'll never understand why MS has this as default behavior. Untick it so you at least have a chance to see the code on screen.
|
United Kingdom20326 Posts
I had auto restart on for ocing 3 systems - it still gives you like 5 seconds at least to see code, in my experience
|
United States1080 Posts
On February 24 2014 16:17 mav451 wrote:
Lol just run it - would have taken you less time to do that than debating on whether to run it at all :p
I already ran BSV, there are no dmp files to read. Took me like 5s to download, 5s to open, and then I was trying to read the documentation to see if I'm missing anything. I don't think I am. My system is set to write to %SystemRoot%\MEMORY.DMP
Are there any locations for dmp files other than windows\minidump\ and windows\memory.dmp ?
C:\Users>echo %SystemRoot%
C:\Windows
Not like my env var is wacked or anything...
On February 24 2014 16:17 mav451 wrote:
Also sounds like you still have it set to Automatically Restart upon failure - I'll never understand why MS has this as default behavior. Untick it so you at least have a chance to see the code on screen.
Will do this, thanks, see if that allows me to catch anything next time I get a BSoD. Not very optimistic about this though, because it seems unusually fast.
On February 24 2014 17:30 Cyro wrote:
I had auto restart on for ocing 3 systems - it still gives you like 5 seconds at least to see code, in my experience
Yea, I've never been unable to capture some useful information before when I had other failures, it's weird. This time I literally don't even have enough time to take a picture when I have my camera all ready to go (maybe I should get a better camera lol).
|
Yeah you shouldn't need to be changing your default settings. I don't know - I've never had a BSOD not give a dump file, so I guess that's where we should focus our efforts.
Hopefully you can catch something on screen at least.
You said there was nothing in EventViewer - are you saying you've had 0 Errors for the lifetime of your machine?
I would assume you at least have the Critical entries "The system has rebooted without cleanly shutting down first," - at the very least. I'd be surprised if it's completely free of Errors or other entries though.
|
United States1080 Posts
On February 25 2014 14:29 mav451 wrote:
Yeah you shouldn't need to be changing your default settings. I don't know - I've never had a BSOD not give a dump file, so I guess that's where we should focus our efforts.
Hopefully you can catch something on screen at least.
Yea, weirdest comp problem I've ever had.
On February 25 2014 14:29 mav451 wrote:You said there was nothing in EventViewer - are you saying you've had 0 Errors for the lifetime of your machine?
I would assume you at least have the Critical entries "The system has rebooted without cleanly shutting down first," - at the very least. I'd be surprised if it's completely free of Errors or other entries though.
I said there was nothing in that time range. There is the expected message about last shutdown not being clean:
On February 24 2014 07:20 phar wrote:I did not disable the page file, have 100GB open space on SSD. Event log from last crash (5:35:51 am according to subsequent system logs) shows nothing, empty for that entire time range Closest event is an hour away, but it's just an uninteresting adobe flash updater turning itself on & off event (occurs every hour).
Just absolutely nothing within 30 minutes of the crash.
|
United States1080 Posts
Well I decided to update my gpu driver (was a year old), and it hasn't crashed since (3-4 days and counting). Because that makes a lot of sense, right.
Give it another week of having random programs open overnight before I give up trying to make it crash again.
|
United States1080 Posts
jinxed it apparently.
This time the program that locked up first was... my taskbar? Windows taskbar disappeared, normal operation for about 15-30 seconds, and then lock & bluescreen. Managed to get a photo of the bluescreen.
Relevant info at the bottom:
STOP: 0x000000F4 (0x0000000000000003, 0xFFFFFA8009F16B30, xFFFFFA8009F16E10, 0xFFFFF800035CB7B0)
Collecting data for crash dump...
Initializing disk for crash dump...
Physical memory dump FAILED with status 0xC0000010.
Maybe related? http://www.techspot.com/community/topics/same-bsod-daily-no-minidump.169741/
I may have to check if my SSD is on a marvel controller.
Will look into stuff more tomorrow,
|
Maybe something from this list is going to help you:
- You should check in Windows system settings->advanced settings->advanced tab->virtual memory settings->advanced->change virtual memory where your page file is currently located. If it is on the ssd change this so it's on the hdd and vice versa. Explanation: code 0xC0000010 means STATUS_INVALID_DEVICE_REQUEST so either the harddrive or/and the ssd is locked up.
- Check on Intel if you have the latest chipset drivers installed.
- Check at Crucial if you have the latest firmware for your ssd.
- Check at harddrive vendor sites if you got the latest firmware.
- Read out Smart data on ssd and harddrive(s) and check for errors.
Good luck
|
United States1080 Posts
Yea I shouldn't need to change anything on my pagefile. It's set up fine for my SSD (where almost everything is), I don't do much on my HDDs (just store pictures) so no need for pagefiles there. I'm not sure what you mean by "if it's on the ssd change this so it's on the hdd, and vice versa". Would this just be for testing purposes? My HDDs are spun down almost all the time, and putting cache on a slow disk sounds like not fun times.
Chipset drivers are updated (at least the most recent ones from my mobo's manufacturer page, though it's from like 1.5 years ago - is intel going to have more updated drivers, and if so how can I make sure they're correct for my mobo?)
I will update my m4's firmware, that's a good idea.
Hard drives are completely unused most of the time (literally, they turn off... they are all fucking insanely loud because they're enterprise drives that I got from work, so I can easily tell when they're on). I seriously doubt there's any problem there. Possible I suppose... will try messing with ssd first.
checked smart data before posting OP, didn't show any errors
|
The error code explanation that Banaora found talks about the drive for the page file not being available to write to when the BSOD occurs. This might mean your SSD is failing in that moment. If that's what's happening, you could put the page file and the mini dump folder location on your HDD so that you get those dumps written when the crash occurs. Both of those things can be configured, the page file location and the mini dump locations.
|
United States1080 Posts
Ah so purely for testing purposes, interesting idea. I think I'd rather just point my minidump to one of my hdds instead of %SystemRoot%\
Thanks, will try and see if that lets me catch a dump file
|
You really also need the page file moved if the BSOD reason is that the SSD froze up. The page file is used as some sort of temporary scratch area or something and is needed to create the dump file. You might also want to switch the crash dump settings over to mini-dump because I don't think there's anything useful in the super large memory.dmp file.
|
United States1080 Posts
Yea already on mini dump. I read up a bit more and can't find anything conclusive, is it really necessary to have a single swap, i.e. disable swap entirely on the ssd, leave swap only on a single hdd to force it to write out a dump? (Is windows not smart enough to check swap on multiple drives before giving up?)
|
United States1080 Posts
Ok got new firmware on my SSD, also determined that it's on an intel controller. Going to hope that windows isn't completely retarded and will let me work with two swap spaces and still be able to save minidumps, but, if that doesn't work on next crash I'll disable the swap space on my ssd. Guess I'll just make sure I don't do too much crazy image processing, and I'll have to change the settings to keep my loud hdds on more often :p
more news at 10 whenever it crashes again randomly
|
United States1080 Posts
|
Have you had any bsods since updating your m4's few?
I had a couple other ideas, but wanted to see if there was any consistency in your bsod codes before moving forward.
|
United States1080 Posts
I have not had a bsod since updating the m4, but it's only been like 4-5 days, so I might just be lucky so far. Unfortunately I'm going away this weekend so I won't be able to leave my computer on overnight to try to catch errors until next week.
Wish I could consistently force the bsod to happen, would make it so much easier to debug.
|
United States1080 Posts
So far no more crashes, so we'll declare this fixed? I dunno, maybe I'll be back in a month after it decides to graduate from bsod to spontaneous combustion.
If not, thanks for all the pointers guys 
|
This means it was caused by the SSD, but there was no way to check this, SMART data was fine etc.?
|
United States1080 Posts
Yea, smart data was fine :\
I've been reading up on all sorts of bad things that crucial m4 firmware can do, but so far nothing that fit exactly with my symptoms.
|
|
|
|
|
|
EPT
