k8s与docker组件堆栈及Debug

k8s 组件日志级别热更新

# 调整日志级别到3
curl -X PUT http://127.0.0.1:8081/debug/flags/v -d "3"

controller manager

wget http://localhost:10252/debug/pprof/profile
wget http://localhost:10252/debug/pprof/heap
curl http://127.0.0.1:10252/debug/pprof/goroutine?debug=1 >> debug1
curl http://127.0.0.1:10252/debug/pprof/goroutine?debug=2 >> debug2

scheduler

kill -12 ${SCHED_PID}
获取scheduler cache信息，输出到日志

kubelet 堆栈信息

wget http://localhost:10250/debug/pprof/profile
wget http://localhost:10250/debug/pprof/heap
curl http://127.0.0.1:10250/debug/pprof/goroutine?debug=1 >> debug1
curl http://127.0.0.1:10250/debug/pprof/goroutine?debug=2 >> debug2

docker 堆栈信息

curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/profile
curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/
curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/

sudo kill -SIGUSR1 $(pidof dockerd)
/var/run/docker/

curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/profile >>docker.profile
curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/goroutine >> docker.goroutine
curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/goroutine?debug=2 >>docker.gorouting_debug_2
curl --unix-socket /var/run/docker.sock -X GET http://v1.2/debug/pprof/heap?debug=2 >>docker.heap

docker-registry 堆栈信息

#先登入机器,然后执行
wget localhost:5002/debug/pprof/profile #这个是cpu占用时间的采样结果，要先等30s
wget localhost:5002/debug/pprof/heap #内存的使用情况
wget localhost:5002/debug/pprof/goroutine?debug=2 #调用栈的全部信息
wget localhost:5002/debug/pprof/goroutine
其他可用的profile:
allocs block goroutine cmdline mutex threadcreate trace，替换上面命令pprof/后面的词即可