 | |
Any help understanding windbg/sos dump of hung app
Dotnet Framework (CLR) microsoft.public.dotnet.framework.clr
 |
| | Thread Tools | Search this Thread | Display Modes |
#1
 
| |||
| |||
|
Any help understanding windbg/sos dump of hung app -
08-10-2006
, 05:41 PM
isablePreemptiveGC
#2
| |||
| |||
|
#3
| |||
| |||
|
|
Wondering if anyone can make better sense of this app dump than I. Working on app that perioldiclly hangs, with maxed out cpu. The app is reading data from hearing aids ( 1 or 2) connected via a serial port to an external device that interfaces w/hearing aid. App is a win forms app written in C#, Dot Net 1.1. Using Windbg I determined that the apps main ui thread is using all the cpu. used adplus to capture app state. Best I can tell is that our main ui thread is in the garbage collecting state and it is acting like it is in a infinite loop. |
code has a busy loop hang and its allocating loads of objects in that
busy loop, chances are that when broken into, it'll be performing a GC.
Quote:
|
Looking at call stack for main ui thread it appears that it had to allocate more memory and a garbage collection was started and it never finished. |
break into it, it happens to be performing a GC?
Quote:
|
0012ed54 0f9787f0 (MethodDesc 0xf7f3bc0 +0x80 System.Data.DataSet.ReadXml), calling (MethodDesc 0xf7f3b90 System.Data.DataSet.ReadXml) 0012ed88 0f9782c5 (MethodDesc 0xf7fedd8 +0x1d5 Starkey.Core.DataServer.Internal.Request..ctor), calling (MethodDesc 0xf7f3bc0 System.Data.DataSet.ReadXml) |
the constructor) calls DataSet.ReadXml and never returns?
-- Barry
--
http://barrkel.blogspot.com/
#4
| |||
| |||
|
|
Hi Jrax, Based on your post, it seems that your application is experiencing a high CPU hang. Can you tell me how do you identify that your main GUI thread is causing the high CPU? Based on the call stack information, it seems that the main GUI thread is hanging in kernel32!WaitForMultipleObjectsEx, which should not cost any CPU time. Can you give it recheck? |
!runaway 3 cmd.
Following is output showing that thread 0 (main ui thread) is only thread
getting cpu time
0:014> !runaway 3
User Mode Time
Thread Time
0:804 0 days 9:51:35.968
2:d8c 0 days 0:00:00.921
12:9dc 0 days 0:00:00.671
7:bc0 0 days 0:00:00.046
9:990 0 days 0:00:00.031
11:704 0 days 0:00:00.015
10:f70 0 days 0:00:00.015
8:c08 0 days 0:00:00.015
14:e8c 0 days 0:00:00.000
13:308 0 days 0:00:00.000
6:878 0 days 0:00:00.000
5:1f8 0 days 0:00:00.000
4:b28 0 days 0:00:00.000
3:fa0 0 days 0:00:00.000
1:ffc 0 days 0:00:00.000
Kernel Mode Time
Thread Time
0:804 0 days 6:31:42.500
12:9dc 0 days 0:00:06.140
2:d8c 0 days 0:00:00.296
7:bc0 0 days 0:00:00.046
5:1f8 0 days 0:00:00.046
9:990 0 days 0:00:00.031
14:e8c 0 days 0:00:00.000
13:308 0 days 0:00:00.000
11:704 0 days 0:00:00.000
10:f70 0 days 0:00:00.000
8:c08 0 days 0:00:00.000
6:878 0 days 0:00:00.000
4:b28 0 days 0:00:00.000
3:fa0 0 days 0:00:00.000
1:ffc 0 days 0:00:00.000
0:014> g
(2dc.dbc): Break instruction exception - code 80000003 (first chance)
eax=7ffd9000 ebx=00000001 ecx=00000002 edx=00000003 esi=00000004 edi=00000005
eip=7c901230 esp=02b4ffcc ebp=02b4fff4 iopl=0 nv up ei pl zr na pe nc
cs=001b ss=0023 ds=0023 es=0023 fs=0038 gs=0000 efl=00000246
ntdll!DbgBreakPoint:
7c901230 cc int 3
0:014> !runaway 3
User Mode Time
Thread Time
0:804 0 days 9:51:43.000
2:d8c 0 days 0:00:00.921
12:9dc 0 days 0:00:00.671
7:bc0 0 days 0:00:00.046
9:990 0 days 0:00:00.031
11:704 0 days 0:00:00.015
10:f70 0 days 0:00:00.015
8:c08 0 days 0:00:00.015
14:dbc 0 days 0:00:00.000
13:308 0 days 0:00:00.000
6:878 0 days 0:00:00.000
5:1f8 0 days 0:00:00.000
4:b28 0 days 0:00:00.000
3:fa0 0 days 0:00:00.000
1:ffc 0 days 0:00:00.000
Kernel Mode Time
Thread Time
0:804 0 days 6:31:46.781
12:9dc 0 days 0:00:06.140
2:d8c 0 days 0:00:00.296
7:bc0 0 days 0:00:00.046
5:1f8 0 days 0:00:00.046
9:990 0 days 0:00:00.031
14:dbc 0 days 0:00:00.000
13:308 0 days 0:00:00.000
11:704 0 days 0:00:00.000
10:f70 0 days 0:00:00.000
8:c08 0 days 0:00:00.000
6:878 0 days 0:00:00.000
4:b28 0 days 0:00:00.000
3:fa0 0 days 0:00:00.000
1:ffc 0 days 0:00:00.000
Quote:
|
The simplest way to identify if a thread is costing high CPU time is using Process Explorer's "Threads" tab in your application's property dialog. You may monitor "CPU" column in this tabpage, it will tell you if the main GUI thread is costing CPU now. Also, it is a good idea to obtain more than 1 dumps from your application and view the main thread call stack, then you can identify if the main GUI thread is really hanging in the mscorwks!Thread::SysSuspendForGC for a long period. |
ui thread is not frozen and I can view disassmbly and step thru code.
Following is some of the method names I see.
KERNEL32!_SEH_prolog:
ntdll!RtlActivateActivationContextUnsafeFast:
KERNEL32!BaseFormatTimeOut:
ntdll!ZwWaitForMultipleObjects:
KERNEL32!BaseSetLastNTError:
ntdll!RtlNtStatusToDosError:
sequence of method calls seem to repeat as I continue to step w/debugger.
Method call sequence leads me to believe that some runtime error is occuring
( _SEH and LastNTError method names ). Anyone know of a register/method
combo I can look at to see what error is occurring.
Also, confused that debugger is not breaking if an error is occuring. Could
there some event filter I need to change?
Quote:
|
Finally, if you require urgent help of analyzing the .Net dump file, I recommend you try to contact Microsoft PSS, they will help you to look into the dump file and perform code review. Thanks. Best regards, Jeffrey Tan Microsoft Online Community Support ================================================== Get notification to my posts through email? Please refer to http://msdn.microsoft.com/subscripti...ult.aspx#notif ications. Note: The MSDN Managed Newsgroup support offering is for non-urgent issues where an initial response from the community or a Microsoft Support Engineer within 1 business day is acceptable. Please note that each follow up response may take approximately 2 business days as the support professional working with you may need further investigation to reach the most efficient resolution. The offering is not appropriate for situations that require urgent, real-time or phone-based interactions or complex project analysis and dump analysis issues. Issues of this nature are best handled working with a dedicated Microsoft Support Engineer by contacting Microsoft Customer Support Services (CSS) at http://msdn.microsoft.com/subscripti...t/default.aspx. ================================================== This posting is provided "AS IS" with no warranties, and confers no rights. |
#5
| ||||
| ||||
|
|
John Aldrin <Jrax (AT) newsgroups (DOT) nospam> wrote: Wondering if anyone can make better sense of this app dump than I. Working on app that perioldiclly hangs, with maxed out cpu. The app is reading data from hearing aids ( 1 or 2) connected via a serial port to an external device that interfaces w/hearing aid. App is a win forms app written in C#, Dot Net 1.1. Using Windbg I determined that the apps main ui thread is using all the cpu. used adplus to capture app state. Best I can tell is that our main ui thread is in the garbage collecting state and it is acting like it is in a infinite loop. Yes, it looks like you caught the code doing a GC. Thing is, if your own code has a busy loop hang and its allocating loads of objects in that busy loop, chances are that when broken into, it'll be performing a GC. |
is repeatedly reading data.
Quote:
|
Looking at call stack for main ui thread it appears that it had to allocate more memory and a garbage collection was started and it never finished. Are you sure about that - and not that it's just that each time you break into it, it happens to be performing a GC? |
thread is only thread getting cpu.
I can view disassembly and step thru code, following is some of the methods
I see getting called.
KERNEL32!_SEH_prolog:
ntdll!RtlActivateActivationContextUnsafeFast:
KERNEL32!BaseFormatTimeOut:
ntdll!ZwWaitForMultipleObjects:
KERNEL32!BaseSetLastNTError:
ntdll!RtlNtStatusToDosError:
Above is not a comprehensive list, but as I keep steping the method names
seem to repeat.
Method call sequence leads me to believe that some runtime error is occuring
( _SEH and LastNTError method names ). Anyone know of a register/method
combo I can look at to see what error is occurring.
Also, confused that debugger is not breaking if an error is occuring. Could
there some event filter I need to change?
Quote:
|
0012ed54 0f9787f0 (MethodDesc 0xf7f3bc0 +0x80 System.Data.DataSet.ReadXml), calling (MethodDesc 0xf7f3b90 System.Data.DataSet.ReadXml) 0012ed88 0f9782c5 (MethodDesc 0xf7fedd8 +0x1d5 Starkey.Core.DataServer.Internal.Request..ctor), calling (MethodDesc 0xf7f3bc0 System.Data.DataSet.ReadXml) Are you certain that Starkey.Core.DataServer.Internal.Request() (i.e. the constructor) calls DataSet.ReadXml and never returns? |
Quote:
|
-- Barry -- http://barrkel.blogspot.com/ |
#6
| |||
| |||
|
|
"Barry Kelly" wrote: John Aldrin <Jrax (AT) newsgroups (DOT) nospam> wrote: Wondering if anyone can make better sense of this app dump than I. Working on app that perioldiclly hangs, with maxed out cpu. The app is reading data from hearing aids ( 1 or 2) connected via a serial port to an external device that interfaces w/hearing aid. App is a win forms app written in C#, Dot Net 1.1. Using Windbg I determined that the apps main ui thread is using all the cpu. used adplus to capture app state. Best I can tell is that our main ui thread is in the garbage collecting state and it is acting like it is in a infinite loop. Yes, it looks like you caught the code doing a GC. Thing is, if your own code has a busy loop hang and its allocating loads of objects in that busy loop, chances are that when broken into, it'll be performing a GC. U are correct. I've written a test app to try to recreate the problem and it is repeatedly reading data. Looking at call stack for main ui thread it appears that it had to allocate more memory and a garbage collection was started and it never finished. Are you sure about that - and not that it's just that each time you break into it, it happens to be performing a GC? I can recreate the problem. I used windbg !runaway 3 to verify that main ui thread is only thread getting cpu. I can view disassembly and step thru code, following is some of the methods I see getting called. KERNEL32!_SEH_prolog: ntdll!RtlActivateActivationContextUnsafeFast: KERNEL32!BaseFormatTimeOut: ntdll!ZwWaitForMultipleObjects: KERNEL32!BaseSetLastNTError: ntdll!RtlNtStatusToDosError: Above is not a comprehensive list, but as I keep steping the method names seem to repeat. Method call sequence leads me to believe that some runtime error is occuring ( _SEH and LastNTError method names ). Anyone know of a register/method combo I can look at to see what error is occurring. Also, confused that debugger is not breaking if an error is occuring. Could there some event filter I need to change? 0012ed54 0f9787f0 (MethodDesc 0xf7f3bc0 +0x80 System.Data.DataSet.ReadXml), calling (MethodDesc 0xf7f3b90 System.Data.DataSet.ReadXml) 0012ed88 0f9782c5 (MethodDesc 0xf7fedd8 +0x1d5 Starkey.Core.DataServer.Internal.Request..ctor), calling (MethodDesc 0xf7f3bc0 System.Data.DataSet.ReadXml) Are you certain that Starkey.Core.DataServer.Internal.Request() (i.e. the constructor) calls DataSet.ReadXml and never returns? I will try to verify. -- Barry -- http://barrkel.blogspot.com/ |
#7
| |||
| |||
|
#8
| |||
| |||
|
|
« Previous Thread
|
Next Thread »
| Thread Tools | Search this Thread |
| Display Modes | |
Linear Mode
| |
Powered by vBulletin Version 3.5.4
Copyright ©2000 - 2009, Jelsoft Enterprises Ltd.
Copyright ©2000 - 2009, Jelsoft Enterprises Ltd.