Contents: Starting special privileged container Operations while running Exiting Persistence How to determine if you're inside the container cleaning up old images Copy/Replace a file into a non-running container

Here are some tips for using docker with Fuego:

Starting [edit section]

After the container is created, you should start it by running:

./start.sh

This calls:

./fuego-host-scripts/docker-start-containter.sh

With no arguments, this will start the container named "fuego-container". However, if you have a multiple fuego containers, or have a container with a different name, you can specify the container name on the start.sh command line.

special privileged container [edit section]

To run the container in a special privileged mode that allows access to host USB devices (needed for accessing Android targets and USB-SERIAL devices), create it using the "--priv" command line option to ./install.sh:

./install.sh --priv

This will call ./fuego-host-scripts/docker-create-usb-privileged-container.sh to create a container with extra privileges to certain host devices.

This script includes a set of devices that the container is granted access to. However, you may need to add a new device to the list. To add a new device to be accessible inside the docker container, please edit the following line in docker-create-usb-privileged-container.sh:

CONTAINER_ID=`sudo docker create -it --privileged -v /dev/bus/usb:/dev/bus/usb -v /dev/ttyACM0:/dev/ttyACM0 ... --net="host" fuego`

With above, only "ttyUSBx" and "ttyACM0" will be detected and accessible inside the docker container.

NOTE: as of February, 2017, this script was in the next branch of the fuego repository.

Operations while running [edit section]

• show the running docker ID
  • sudo docker ps
• execute a command in the container
  • docker exec <id> <some_command>
• attach another shell inside the container
  • sudo docker exec -i -t <id> bash
  • there is a helper script called 'fuegosh' which can be used to get a shell inside a currently running Fuego container
    • see fuego-core/scripts/fuegosh
• access docker container using ssh
  • ssh user@<ip_addr> -p 2222
  • sshd is running on 2222 in the container, if the default sshd_config is used
• copy files to the container
  • docker cp foo <id>:/path/to/dest
• copy files from the container
  • docker cp <id>:/path/to/src/foo bar

Exiting [edit section]

To exit the docker container, just exit the primary shell that started with the container was started.

Persistence [edit section]

The Fuego container uses docker bind mounts so that some files persist in the host filesystem, even when the container is not running.

In the host system, these are under fuego-ro, fuego-rw and fuego-core in the directory where the container was created.

Here are some files that persist:

• fuego-ro/boards* - for board definition files
• fuego-ro/conf/ttc.conf - for use with ttc targets
• fuego-ro/toolchains - this is where toolchains and SDKs can be installed
• fuego-ro/toolchains/tools.sh - this file has the multiplexor for the different toolchains (on the PLATFORM variable)
• fuego-rw/logs - this has logs from executed test runs
• fuego-rw/work
• fuego-rw/buildzone - this is where test programs are built
• fuego-rw/test - place where the board 'docker' places test materials

How to determine if you're inside the container [edit section]

• grep -q docker /proc/1/cgroup ; echo $?
  • will be 0 if inside the container, 1 if on host

cleaning up old images [edit section]

I build lots of docker images, and they leave lots of data around.

• docker ps -a - show docker containers on your system, and their images
• docker images - show images on your system, and their age and size
• docker rmi <id> - remove an image (you must remove any containers using this image first)
• docker rm <id> - remove a container

Copy/Replace a file into a non-running container [edit section]

Background: Consider a case where you make some changes to '/etc/default/jenkins' file when your container is running, and then you restart the container. Unfortunately your container may not start because of an issue in the '/etc/default/jenkins' file. How do you fix it as the container itself is not running?

Solution: Get the container id (of the non-running container) via

$ 'docker ps -a' command

Replace the faulty file with original/corrected one via 'docker cp' command as shown in the example below.

$ sudo docker cp jenkins 6b4e6e63rfg7:/etc/default/

where '6b4e6e63rfg7' is the container id of the non-running container

Now you will able to start the docker container successfully.