Error starting container with balena-socket label

Product support
1

Hello, I am interested in launching a container with the Balena (Docker) socket mounted inside. I know there is a label for this, yet when I try to start the container w/ this configuration, I get this error:

Error response from daemon: OCI runtime create failed: container_linux.go:349: starting container process caused "process_linux.go:449: container init caused \"rootfs_linux.go:58: mounting \\\"/var/run/balena-engine.sock\\\" to rootfs \\\"/var/lib/docker/aufs/mnt/756dcc99fbbf574084de617ece9a4347198b7232bf7d83371f4d3bc20a539078\\\" at \\\"/host/run/balena-engine.sock\\\" caused \\\"stat /var/run/balena-engine.sock: no such file or directory\\\"\"": unknown

The /var/run/balena-engine.sock file is on the host OS, so I’m not sure what this is complaining about exactly. Is this possibly a bug?

My fleet is using a docker-compose file, which looks like this:

---
version: "2.1"

services:
  k3s:
    build: ./k3s
    container_name: k3s
    tmpfs:
      - /run
      - /var/run
    ulimits:
      nproc: 65535
      nofile:
        soft: 65535
        hard: 65535
    privileged: true
    network_mode: host
    restart: always
    volumes:
      - rancher_etc:/etc/rancher
    environment:
      - K3S_URL=${K3S_URL}
      - K3S_TOKEN=${K3S_TOKEN:?err}
    labels:
      io.balena.features.balena-socket: '1'
      io.balena.features.kernel-modules: '1'
      io.balena.features.firmware: '1'
      io.balena.features.sysfs: '1'

Update: writing this made me reflect on the fact that my fleet is specifying /var/run as a tmpfs mount, which is probably the problem. Will verify that this goes away if I remove that.

2

Solved my own problem – it was indeed because I was specifying /var/run as a tmpfs mount, which conflicted w/ the bind mount that is created due to the label.