Extreme Memory Leak: Almost 100,000 Handles a Day. MEMORY_MANAGEMENT STOP Code
General discussion
-
Hello,
I am experiencing a HUGE memory leak working on Windows Developer Preview.
According to task manager, the leak is about 100,000 handles a day. The OS stayed well @ over ENORMOUS 750,000 handles for 19 days and 20 hours and ended up with MEMORY_MANAGEMENT STOP code.
The Task Manager shows the following:
I used PowerShell and that's what I got:
PS C:\Windows\system32> Get-Process | Where-Object {$_.handlecount -gt 10000}
Handles NPM(K) PM(K) WS(K) VM(M) CPU(s) Id ProcessName
------- ------ ----- ----- ----- ------ -- -----------
2568050 59 139204 539324 874 1 019,29 868 dwm
Two and a half million handles for Desktop Window Manager!
I do NOT have the kernel debugger so I can't just kd !analyze here.
The OS created a minidump in %windir%\minidump.
I overlooked the Action Center's settings and it showed the following:
Do I understand it like OS has already sent the minidump to bugtracking servers? Or do I have to do anything to make it done?
Additional:
1. The OS natively boots in VHD.
2. The OS is using nVidia video drivers and VIA audio drivers.
3. In the System event log I see the following records:
The previous system shutdown at 11:41:23 AM on 10/13/2011 was unexpected.The computer has rebooted from a bugcheck. The bugcheck was: 0x0000001a (0x0000000000041790, 0xfffffa80064fecf0, 0x0000000000000000, 0x0000000000000000). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: 101311-121056-01.
An I/O operation initiated by the Registry failed unrecoverably.The Registry could not flush hive (file): ''.
The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.
The following boot-start or system-start driver(s) failed to load:3ware
adp94xx
adpahci
adpu320
amdsata
amdsbs
amdxata
arc
arcsas
b06bdrv
bfad
dam
ebdrv
EhStorTcgDrv
elxfcoe
elxstor
HpSAMD
iaStorV
iirsp
intelide
isapnp
LSI_FC
LSI_SAS
LSI_SAS2
LSI_SCSI
LSI_SSS
megasas
MegaSR
mvumis
nfrd960
nvraid
nvstor
ql2300
ql40xx
sbp2port
SiSRaid2
SiSRaid4
stexstor
storahci
storflt
storvsc
viaide
vmbus
vsmraid
VSTXRAID
Wd
WdBootAll the listed errors occurred after the STOP error.
I understand that this is even not a beta and that it is normal to face the problems as such. The only reason I am asking is I want to help MSFT to figure out the possible root cause of the problem, that's it. If I knew how I could help, I would do my best to collect necessary info.
I will be glad for any comments. Thank you.
BTW: the new Task Manager is simply AWESOME!
Well this is the world we live in And these are the hands we're given...
- Edited by Exotic Hadron Friday, October 14, 2011 10:12 AM
All replies
-
Thank you for your reply, Steven!
The problem is that the Reliability Monitor DOES NOT show if it has sent the dump or not...
Problem on the problem calendar:
Description of the problem:
I would love to send the dump file. Do I understand it right that I will be able to send it using the Windows Feedback Tools? Correct?
Or is there any other way I could do that?Okay, I've got the Feedback Tool installed. How do I submit information about the fault?
On the Identify your issue page I see Select area. What area do I have to choose to report a crash? I don't know what crash actually. I SUPPOSE it's because of memory leak.
And I also know that I had
STOP Error 0x0000001A: MEMORY_MANAGEMENT
STOP error 0x1A means that a severe memory management error occurred. STOP code 0x0000001A may also display "MEMORY_MANAGEMENT" on the same STOP message.
And I suppose that this is caused by a leak in WDM process.
Steven, could you please guide me through here?
Well this is the world we live in And these are the hands we're given...
- Edited by Exotic Hadron Tuesday, October 18, 2011 7:46 AM
-
I am still trying to find the reproduce steps, but it is pretty impressive that the computer did not slow down until the handle count of DWM exceeds 1.2 million.
The following is signature, not part of post
Please mark the post answered your question as the answer, and mark other helpful posts as helpful, so they will appear differently to other users who are visiting your thread for the same problem.
Visual C++ MVP
- Edited by Sheng Jiang 蒋晟 Friday, October 28, 2011 1:40 AM
-
I am still trying to find the reproduce steps, but it is pretty impressive that the computer runs normally until the handle count of DWM exceeds 1.2 million.
The following is signature, not part of post
Please mark the post answered your question as the answer, and mark other helpful posts as helpful, so they will appear differently to other users who are visiting your thread for the same problem.
Visual C++ MVP- Edited by Sheng Jiang 蒋晟 Friday, October 28, 2011 1:29 AM
