jangocheng / lldb-netcore

lldb-netcore

Docker image with lldb debugger and SOS plugin, compiled from sources with lldb headers. By default loads process coredump from /tmp/coredump, loads SOS plugin and prints current exception, leaving lldb shell open. Image tag matches dotnet runtime version.

How to use

docker run --rm -it -v /stripe/upload/coredump:/tmp/coredump 6opuc/lldb-netcore

• /stripe/upload/coredump - Path to coredump of crashed process on docker host machine

Usecases

Container crashed

1. Copy crashed process working directory(coredump is automatically created in crashed process working directory) to temporary directory on host:

docker cp 79686a7aff63:/app /tmp

• 79686a7aff63 - id of container with crashed process
• /app - crashed process working directory inside container filesystem

2. Find crashed process coredump:

ls /tmp/app/core.*

example output:

/tmp/app/core.26939

3. Open coredump with debugger:

docker run --rm -it -v /tmp/app/core.26939:/tmp/coredump 6opuc/lldb-netcore

example output:

(lldb) target create "/usr/bin/dotnet" --core "/tmp/coredump"
Core file '/tmp/coredump' (x86_64) was loaded.
(lldb) plugin load /coreclr/libsosplugin.so
(lldb) sos PrintException
Exception object: 00007f3fb001ce08
Exception type:   System.NullReferenceException
Message:          Object reference not set to an instance of an object.
InnerException:   <none>
StackTrace (generated):
    SP               IP               Function
    00007FFCE0A312F0 00007F3FD7940481 test.dll!test.Program.Main(System.String[])+0x1
StackTraceString: <none>
HResult: 80004003
(lldb)

4. Continue exploring coredump in lldb shell:

help

Analyze running container

1. Get id of docker container(docker ps) you need to analyze. In this example it is "b5063ef5787c"

2. Run container with createdump utility(it needs sys_admin and sys_ptrace privileges. If your running container already has these privileges you can attach to running container and run createdump utility from there):

docker run --rm -it --cap-add sys_admin --cap-add sys_ptrace --net=container:b5063ef5787c --pid=container:b5063ef5787c -v /tmp:/tmp 6opuc/lldb-netcore /bin/bash

• b5063ef5787c - id of container you need to analyze
• /tmp - temporary directory on host, where coredump will be created

3. Find PID of dotnet process you need to analyze:

ps aux

In this example PID is "7"

4. Create coredump of dotnet process and exit from container:

createdump -u -f /tmp/coredump 7
exit

• 7 is dotnet process PID

5. Open coredump with debugger:

docker run --rm -it -v /tmp/coredump:/tmp/coredump 6opuc/lldb-netcore

example output:

(lldb) target create "/usr/bin/dotnet" --core "/tmp/coredump"
Core file '/tmp/coredump' (x86_64) was loaded.
(lldb) plugin load /coreclr/libsosplugin.so
(lldb) sos PrintException
There is no current managed exception on this thread
(lldb)

6. Continue exploring coredump in lldb shell:

help

How to build

netcore 2.2.3:

docker build \
	--tag 6opuc/lldb-netcore:2.2.3 \
   	--build-arg BASE_IMAGE=mcr.microsoft.com/dotnet/core/sdk:2.2.105 \
   	--build-arg CORECLR_BRANCH=v2.2.3 \
	.

netcore 2.2.2:

docker build \
	--tag 6opuc/lldb-netcore:2.2.2 \
   	--build-arg BASE_IMAGE=mcr.microsoft.com/dotnet/core/sdk:2.2.104 \
   	--build-arg CORECLR_BRANCH=v2.2.2 \
	.

netcore 2.1.8:

docker build \
	--tag 6opuc/lldb-netcore:2.1.8 \
	--build-arg BASE_IMAGE=mcr.microsoft.com/dotnet/core/sdk:2.1.504 \
   	--build-arg CORECLR_BRANCH=v2.1.8 \
	.

netcore 2.0.3:

docker build \
	--tag 6opuc/lldb-netcore:2.0.3 \
	--build-arg BASE_IMAGE=microsoft/dotnet:2.0.3-sdk \
	--build-arg CORECLR_BRANCH=v2.0.3 \
	.

• BASE_IMAGE - Base image of dotnet sdk. Used both at build time and runtime.
• CORECLR_BRANCH - coreclr repository(https://github.com/dotnet/coreclr.git) branch/tag to build SOS plugin from