Debug a deadlock in .NET Core
This article applies to: ✔️ .NET Core 3.1 SDK and later versions
In this tutorial, you'll learn how to debug a deadlock scenario. Using the provided example ASP.NET Core web app source code repository, you can cause a deadlock intentionally. The endpoint will stop responding and experience thread accumulation. You'll learn how you can use various tools to analyze the problem, such as core dumps, core dump analysis, and process tracing.
In this tutorial, you will:
- Investigate an app that has stopped responding
- Generate a core dump file
- Analyze process threads in the dump file
- Analyze call stacks and sync blocks
- Diagnose and solve a deadlock
Prerequisites
The tutorial uses:
- .NET Core 3.1 SDK or a later version
- Sample debug target - web app to trigger the scenario
- dotnet-trace to list processes
- dotnet-dump to collect, and analyze a dump file
Core dump generation
To investigate application unresponsiveness, a core dump or memory dump allows you to inspect the state of its threads and any possible locks that may have contention issues. Run the sample debug application using the following command from the sample root directory:
dotnet run
To find the process ID, use the following command:
dotnet-trace ps
Take note of the process ID from your command output. Our process ID was 4807, but yours will be different. Navigate to the following URL, which is an API endpoint on the sample site:
https://localhost:5001/api/diagscenario/deadlock
The API request to the site will stop responding. Let the request run for about 10-15 seconds. Then create the core dump using the following command:
Analyze the core dump
To start the core dump analysis, open the core dump using the following dotnet-dump analyze command. The argument is the path to the core dump file that was collected earlier.
dotnet-dump analyze ~/.dotnet/tools/core_20190513_143916
Since you're looking at a potentially unresponsive application, you want an overall feel for the thread activity in the process. You can use the threads command as shown below:
> threads
*0 0x1DBFF (121855)
1 0x1DC01 (121857)
2 0x1DC02 (121858)
3 0x1DC03 (121859)
4 0x1DC04 (121860)
5 0x1DC05 (121861)
6 0x1DC06 (121862)
7 0x1DC07 (121863)
8 0x1DC08 (121864)
9 0x1DC09 (121865)
10 0x1DC0A (121866)
11 0x1DC0D (121869)
12 0x1DC0E (121870)
13 0x1DC10 (121872)
14 0x1DC11 (121873)
15 0x1DC12 (121874)
16 0x1DC13 (121875)
17 0x1DC14 (121876)
18 0x1DC15 (121877)
19 0x1DC1C (121884)
20 0x1DC1D (121885)
21 0x1DC1E (121886)
22 0x1DC21 (121889)
23 0x1DC22 (121890)
24 0x1DC23 (121891)
25 0x1DC24 (121892)
26 0x1DC25 (121893)
...
...
317 0x1DD48 (122184)
318 0x1DD49 (122185)
319 0x1DD4A (122186)
320 0x1DD4B (122187)
321 0x1DD4C (122188)
The output shows all the threads currently running in the process with their associated debugger thread ID and operating system thread ID. Based on the output, there are over 300 threads.
The next step is to get a better understanding of what the threads are currently doing by getting each thread's callstack. The clrstack command can be used to output callstacks. It can either output a single callstack or all the callstacks. Use the following command to output all the callstacks for all the threads in the process:
clrstack -all
A representative portion of the output looks like:
...
...
...
Child SP IP Call Site
00007F2AE37B5680 00007f305abc6360 [GCFrame: 00007f2ae37b5680]
00007F2AE37B5770 00007f305abc6360 [GCFrame: 00007f2ae37b5770]
00007F2AE37B57D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae37b57d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE37B5920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE37B5950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE37B5CA0 00007f30593044af [GCFrame: 00007f2ae37b5ca0]
00007F2AE37B5D70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae37b5d70]
OS Thread Id: 0x1dc82
Child SP IP Call Site
00007F2AE2FB4680 00007f305abc6360 [GCFrame: 00007f2ae2fb4680]
00007F2AE2FB4770 00007f305abc6360 [GCFrame: 00007f2ae2fb4770]
00007F2AE2FB47D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae2fb47d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE2FB4920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE2FB4950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE2FB4CA0 00007f30593044af [GCFrame: 00007f2ae2fb4ca0]
00007F2AE2FB4D70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae2fb4d70]
OS Thread Id: 0x1dc83
Child SP IP Call Site
00007F2AE27B3680 00007f305abc6360 [GCFrame: 00007f2ae27b3680]
00007F2AE27B3770 00007f305abc6360 [GCFrame: 00007f2ae27b3770]
00007F2AE27B37D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae27b37d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE27B3920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE27B3950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE27B3CA0 00007f30593044af [GCFrame: 00007f2ae27b3ca0]
00007F2AE27B3D70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae27b3d70]
OS Thread Id: 0x1dc84
Child SP IP Call Site
00007F2AE1FB2680 00007f305abc6360 [GCFrame: 00007f2ae1fb2680]
00007F2AE1FB2770 00007f305abc6360 [GCFrame: 00007f2ae1fb2770]
00007F2AE1FB27D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae1fb27d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE1FB2920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE1FB2950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE1FB2CA0 00007f30593044af [GCFrame: 00007f2ae1fb2ca0]
00007F2AE1FB2D70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae1fb2d70]
OS Thread Id: 0x1dc85
Child SP IP Call Site
00007F2AE17B1680 00007f305abc6360 [GCFrame: 00007f2ae17b1680]
00007F2AE17B1770 00007f305abc6360 [GCFrame: 00007f2ae17b1770]
00007F2AE17B17D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae17b17d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE17B1920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE17B1950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE17B1CA0 00007f30593044af [GCFrame: 00007f2ae17b1ca0]
00007F2AE17B1D70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae17b1d70]
OS Thread Id: 0x1dc86
Child SP IP Call Site
00007F2AE0FB0680 00007f305abc6360 [GCFrame: 00007f2ae0fb0680]
00007F2AE0FB0770 00007f305abc6360 [GCFrame: 00007f2ae0fb0770]
00007F2AE0FB07D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae0fb07d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE0FB0920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE0FB0950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE0FB0CA0 00007f30593044af [GCFrame: 00007f2ae0fb0ca0]
00007F2AE0FB0D70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae0fb0d70]
OS Thread Id: 0x1dc87
Child SP IP Call Site
00007F2AE07AF680 00007f305abc6360 [GCFrame: 00007f2ae07af680]
00007F2AE07AF770 00007f305abc6360 [GCFrame: 00007f2ae07af770]
00007F2AE07AF7D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2ae07af7d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2AE07AF920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2AE07AF950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2AE07AFCA0 00007f30593044af [GCFrame: 00007f2ae07afca0]
00007F2AE07AFD70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2ae07afd70]
OS Thread Id: 0x1dc88
Child SP IP Call Site
00007F2ADFFAE680 00007f305abc6360 [GCFrame: 00007f2adffae680]
00007F2ADFFAE770 00007f305abc6360 [GCFrame: 00007f2adffae770]
00007F2ADFFAE7D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2adffae7d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2ADFFAE920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2ADFFAE950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2ADFFAECA0 00007f30593044af [GCFrame: 00007f2adffaeca0]
00007F2ADFFAED70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2adffaed70]
...
...
Observing the callstacks for all 300+ threads shows a pattern where a majority of the threads share a common callstack:
OS Thread Id: 0x1dc88
Child SP IP Call Site
00007F2ADFFAE680 00007f305abc6360 [GCFrame: 00007f2adffae680]
00007F2ADFFAE770 00007f305abc6360 [GCFrame: 00007f2adffae770]
00007F2ADFFAE7D0 00007f305abc6360 [HelperMethodFrame_1OBJ: 00007f2adffae7d0] System.Threading.Monitor.ReliableEnter(System.Object, Boolean ByRef)
00007F2ADFFAE920 00007F2FE392B31F testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_1() [/home/marioh/webapi/Controllers/diagscenario.cs @ 36]
00007F2ADFFAE950 00007F2FE392B46D System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/__w/3/s/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
00007F2ADFFAECA0 00007f30593044af [GCFrame: 00007f2adffaeca0]
00007F2ADFFAED70 00007f30593044af [DebuggerU2MCatchHandlerFrame: 00007f2adffaed70]
The callstack seems to show that the request arrived in our deadlock method that in turn makes a call to Monitor.ReliableEnter. This method indicates that the threads are trying to enter a monitor lock. They're waiting on the availability of the lock. It's likely locked by a different thread.
The next step then is to find out which thread is actually holding the monitor lock. Since monitors typically store lock information in the sync block table, we can use the syncblk command to get more information:
> syncblk
Index SyncBlock MonitorHeld Recursion Owning Thread Info SyncBlock Owner
43 00000246E51268B8 603 1 0000024B713F4E30 5634 28 00000249654b14c0 System.Object
44 00000246E5126908 3 1 0000024B713F47E0 51d4 29 00000249654b14d8 System.Object
-----------------------------
Total 344
CCW 1
RCW 2
ComClassFactory 0
Free 0
The two interesting columns are MonitorHeld and Owning Thread Info. The MonitorHeld column shows whether a monitor lock is acquired by a thread and the number of waiting threads. The Owning Thread Info column shows which thread currently owns the monitor lock. The thread info has three different subcolumns. The second subcolumn shows operating system thread ID.
At this point, we know two different threads (0x5634 and 0x51d4) hold a monitor lock. The next step is to take a look at what those threads are doing. We need to check if they're stuck indefinitely holding the lock. Let's use the setthread and clrstack commands to switch to each of the threads and display the callstacks.
To look at the first thread, run the setthread command, and find the index of the 0x5634 thread (our index was 28). The deadlock function is waiting to acquire a lock, but it already owns the lock. It's in deadlock waiting for the lock it already holds.
> setthread 28
> clrstack
OS Thread Id: 0x5634 (28)
Child SP IP Call Site
0000004E46AFEAA8 00007fff43a5cbc4 [GCFrame: 0000004e46afeaa8]
0000004E46AFEC18 00007fff43a5cbc4 [GCFrame: 0000004e46afec18]
0000004E46AFEC68 00007fff43a5cbc4 [HelperMethodFrame_1OBJ: 0000004e46afec68] System.Threading.Monitor.Enter(System.Object)
0000004E46AFEDC0 00007FFE5EAF9C80 testwebapi.Controllers.DiagScenarioController.DeadlockFunc() [C:\Users\dapine\Downloads\Diagnostic_scenarios_sample_debug_target\Controllers\DiagnosticScenarios.cs @ 58]
0000004E46AFEE30 00007FFE5EAF9B8C testwebapi.Controllers.DiagScenarioController.<deadlock>b__3_0() [C:\Users\dapine\Downloads\Diagnostic_scenarios_sample_debug_target\Controllers\DiagnosticScenarios.cs @ 26]
0000004E46AFEE80 00007FFEBB251A5B System.Threading.ThreadHelper.ThreadStart_Context(System.Object) [/_/src/System.Private.CoreLib/src/System/Threading/Thread.CoreCLR.cs @ 44]
0000004E46AFEEB0 00007FFE5EAEEEEC System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object) [/_/src/System.Private.CoreLib/shared/System/Threading/ExecutionContext.cs @ 201]
0000004E46AFEF20 00007FFEBB234EAB System.Threading.ThreadHelper.ThreadStart() [/_/src/System.Private.CoreLib/src/System/Threading/Thread.CoreCLR.cs @ 93]
0000004E46AFF138 00007ffebdcc6b63 [GCFrame: 0000004e46aff138]
0000004E46AFF3A0 00007ffebdcc6b63 [DebuggerU2MCatchHandlerFrame: 0000004e46aff3a0]
The second thread is similar. It's also trying to acquire a lock that it already owns. The remaining 300+ threads that are all waiting are most likely also waiting on one of the locks that caused the deadlock.
See also
- dotnet-trace to list processes
- dotnet-counters to check managed memory usage
- dotnet-dump to collect and analyze a dump file
- dotnet/diagnostics
Next steps
Feedback
Coming soon: Throughout 2024 we will be phasing out GitHub Issues as the feedback mechanism for content and replacing it with a new feedback system. For more information see: https://aka.ms/ContentUserFeedback.
Submit and view feedback for