.net core在linux docker部署，内存异常分析，分析工具：dotnet-dump

老项目为.net 4.6，部署在iis，设置的32位，消耗内存基本稳定在500mb左右，但是升级为.net 6，部署在linux docker上以后，内存消耗异常，记录一下

docker stats查看，用户模块占10g

一,创建dump文件

注：执行/usr/share/dotnet/shared/Microsoft.NETCore.App/6.0.25/createdump 1 这句的时候，会导致程序无法访问，内存越大，保存时间越长，请选择合适的时间进行此操作。

[root@WEB1 zwsec]# docker exec -it base bash
root@68a12d966fa4:/src# find / -name createdump
/usr/share/dotnet/shared/Microsoft.NETCore.App/6.0.25/createdump
root@68a12d966fa4:/src# /usr/share/dotnet/shared/Microsoft.NETCore.App/6.0.25/createdump 1
[createdump] Gathering state for process 1 dotnet
[createdump] Writing minidump with heap to file /tmp/coredump.1
[createdump] Written 13254832128 bytes (3236043 pages) to core file
[createdump] Target process is alive
[createdump] Dump successfully written
root@68a12d966fa4:/src# exit
exit
[root@WEB1 zwsec]# docker cp base:/tmp/coredump.1 /data1/app/server/test_dump/coredump.1
[root@WEB1 zwsec]# cd /data1/app/server/test_dump/
[root@WEB1 test_dump]# ls
coredump.1

二、分析dump文件

dotnet-dump工具依赖dotnet的sdk，如果宿主机中安装了sdk可以直接在宿主机中分析；
如果不想污染宿主机环境可以拉取一个sdk镜像mcr.microsoft.com/dotnet/sdk，创建一个临时环境用于分析。

1,创建一个用于分析的临时容器

需要把用于存放coredump.1文件的目录挂载到容器，或者自己cp进去。
创建好容器之后需要安装dotnet-dump工具

docker run --security-opt seccomp=unconfined --name test --rm -it -v /data1/app/server/test_dump:/tmp/coredump mcr.microsoft.com/dotnet/sdk

注：需要加上 --security-opt seccomp=unconfined 否则安装dotnet-dump时可能报错

[root@WEB1 zwsec]# docker run --security-opt seccomp=unconfined --name test --rm -it -v /data1/app/server/test_dump:/tmp/coredump mcr.microsoft.com/dotnet/sdk
Unable to find image 'mcr.microsoft.com/dotnet/sdk:latest' locally
latest: Pulling from dotnet/sdk
fd674058ff8f: Pull complete
d92ceca2d7b5: Pull complete
408b46bb4d4d: Pull complete
e9dfe0944a0c: Pull complete
3425c4117635: Pull complete
e4513f6c2cfb: Pull complete
5e118823fce1: Pull complete
66fb1dfc2c9d: Pull complete
683a16faa7de: Pull complete
Digest: sha256:3fcf6f1e809c0553f9feb222369f58749af314af6f063f389cbd2f913b4ad556
Status: Downloaded newer image for mcr.microsoft.com/dotnet/sdk:latest
root@d99cf52da4ed:/# cd /tmp/coredump
root@d99cf52da4ed:/tmp/coredump# ls
coredump.1
root@d99cf52da4ed:/tmp/coredump# dotnet tool install -g dotnet-dump
Tools directory '/root/.dotnet/tools' is not currently on the PATH environment variable.
If you are using bash, you can add it to your profile by running the following command:

cat << \EOF >> ~/.bash_profile
# Add .NET Core SDK tools
export PATH="$PATH:/root/.dotnet/tools"
EOF

You can add it to the current session by running the following command:

export PATH="$PATH:/root/.dotnet/tools"

You can invoke the tool using the following command: dotnet-dump
Tool 'dotnet-dump' (version '9.0.553101') was successfully installed.
root@d99cf52da4ed:/tmp/coredump# export PATH="$PATH:/root/.dotnet/tools"
root@d99cf52da4ed:/tmp/coredump#

2,分析内存泄漏

命令：

1.dotnet-dump analyze coredump.1

2.dumpheap -stat 找到堆上的对象信息（注：此命令第一次执行时会消耗很久时间，请泡一杯茶耐心等候，如果服务器会自动断开连接的，建议下载到一台不会断开连接的机器进行分析，因为这里可能在没有分析出来连接就断开了）

3.dumpheap -mt <mt> 列出所有与<mt>结构对应的对象，一般我都会找string对象的mt，看看里面大size的string对象是有哪些

4.do <address> 查看对象详细信息

5.gcroot -all <address> 一般我也是找string地址，看看对象引用

第1，2个命令结果：（请耐心等待，如果文件很大，请使用一台靠谱的机器分析）

root@d99cf52da4ed:/tmp/coredump# dotnet-dump analyze coredump.1
Loading core dump: coredump.1 ...
Ready to process analysis commands. Type 'help' to list available commands or 'help [command]' to get detailed help on a command.
Type 'quit' or 'exit' to exit the session.
> dumpheap -stat

Statistics: MT Count TotalSize Class Name

…………

7f19789cbdd0 808,824 71,176,512 System.RuntimeMethodInfoStub
7f197e149d88 758,432 78,876,928 Newtonsoft.Json.Linq.JProperty
7f197cffb878 7,107,012 170,568,288 System.WeakReference<Microsoft.Extensions.DependencyInjection.ServiceProvider>
7f1977244ab8 350,296 179,511,302 System.Char[]
7f19789f5598 1,818,443 307,396,110 System.Byte[]
7f197d74f1f0 325,913 433,410,924 <unknown_type_7f197d74f1f0>
7f19771bd2e0 4,481,237 626,510,840 System.String
563ee4af1ea0 497,037 1,051,396,144 Free
Total 33,543,157 objects, 4,234,311,341 bytes