system

Assembly of the complete system in the form of both an Ansible script and as a Docker image

Ansible

Setup

ansible-galaxy install -r install_roles.yml -p roles

Testing the script with Vagrant

vagrant up
make vagrant
ansible-playbook test-server.yml -l test

Now open the web UI at http://192.168.50.4:9000. When asked for the web API's address for the Pipeline 2 Engine, enter "http://localhost:8181/ws".

Docker image

Building the image

It is recommended using the pre-built images available at hub.docker.com/r/snaekobbi/system/. Use the image ID snaekobbi/system to get the latest version of the system, and optionally add a tag to get a specific version of the system, for instance snaekobbi/system:v1.7.0.

If instead you want to build the image yourself, you can run this command:

docker build .

The last line will display the id of the image that was built. Something like:

Successfully built 72ca1a0390e8

In this case, 72ca1a0390e8 is the image ID.

Basic usage with both Web UI and Engine

docker run -it -p 9000:9000 -p 8181:8181 snaekobbi/system:1.7.0-latest

Now, open http://localhost:9000/ in a browser.

If you are not running GNU/Linux, you will have to use the IP address of your VM, which you can find by running docker-machine ls. An alternative is to use port forwarding. For VirtualBox driver: VBoxManage controlvm <name of docker machine> natpf1 "tcp-port9000,tcp,127.0.0.1,9000,,9000".

You can also query the engine directly at the endpoint http://localhost:8181/ws/. For instance, try the following URLs in a browser (refer to the Web API docs for more information):

• http://localhost:8181/ws/alive
• http://localhost:8181/ws/scripts
• http://localhost:8181/ws/jobs

Running an interactive shell

Appending bash to the end will start an interactive shell instead of starting the engine+webui:

docker run -it -p 9000:9000 -p 8181:8181 snaekobbi/system:1.7.0-latest bash

In the shell you can use the dp2 command to interact with Pipeline 2. The dp2 command will start the engine if it's not running already.

To start the engine and webui manually:

service pipeline2d start
service daisy-pipeline2-webui start

Running multiple conversions as a batch job

There are several ways to do this of course, this is just an example.

• Create a directory with the input books: $HOME/snaekobbi/system/target/input
• Put all your DTBooks there (551848.xml, 554569.xml, etc.)
• Create an empty directory for the output books: $HOME/snaekobbi/system/target/output
• Create a bash script that creates jobs, waits for them to finish, and optionally creates logs: $HOME/snaekobbi/system/target/batch.sh
  • Here's an example: https://gist.github.com/josteinaj/efa1ca9be1ebce5517d8a91b1911e682

To run this batch conversion, do something like this:

docker run -it \
        -v $HOME/snaekobbi/system/target/input:/mnt/input:ro \
        -v $HOME/snaekobbi/system/target/output:/mnt/output \
        -v $HOME/snaekobbi/system/target/batch.sh:/mnt/script/batch.sh \
        snaekobbi/system:1.7.0-latest /mnt/script/batch.sh

Release procedure

• Create a release branch.

  git checkout -b release/${VERSION}

• Resolve snapshot versions in roles/test-server/vars/debs.yml and commit.

• Make release notes and commit.

• Tag

  git tag -s -a v${VERSION} -m "Version ${VERSION}"

• Push the tag.

  git push origin v${VERSION}

• Add the release notes to http://github.com/snaekobbi/system/releases/v${VERSION}.