Shutting down MongoDB inside Docker container causes 100% CPU

I’ve been pulling my hair over this one for days now, to the point I am dropping this here. Using docker-compose I start a few containers, one of which is running mongodb, using CMD [“/usr/bin/mongod”] However, when I shut down the container using docker stop, the mongod process takes 100% cpu and keeps running forever without being able to kill it. When killing the parent processes, the mongod process becomes a child of init, but in its same state. Mongodb does log that it shutted down with error code 0 though.

Docker version: 17.04.0-ce
Docker-compose version: 1.11.2

  • I start docker with a shell script but when the shell script exits, the container crashes
  • debug docker deamon with gdb or cgdb
  • docker command attach running slow
  • Testing a Node library working with Docker containers
  • Docker: how to send a signal from one running container to another one?
  • Unable to mount specific paths as volumes
  • What I tried so far:

    • different linux distro’s
    • supervisord within container to handle mongod
    • added –init to docker daemon
    • run mongod as a fork
    • look for kernel update, but 3.14 seems to be the latest (running on Odroid C2)
    • different storage drivers, but only aufs works (overlay needs higher kernel)

    All without any joy..

    enter image description here

  • Timeout issue on while running docker run for onlyoffice document server integration
  • Dokku/Docker, how to access file in file system of running container?
  • Canonical way to checksum downloads in a Dockerfile?
  • How to pass parameters to a .net core project with dockerfile
  • docker login in docker machine
  • Pulling Docker Images from Private Repository using REGISTRY REST API
  • One Solution collect form web for “Shutting down MongoDB inside Docker container causes 100% CPU”

    After updating the kernel to 4.x on the Odroid-C2 using this link, the problem went away.

    Docker will be the best open platform for developers and sysadmins to build, ship, and run distributed applications.