A Docker image to automate Cloud Foundry deployments
This alpine based docker image uses the Cloud Foundry CLI in version 7.1.0 to carry out deployments.
cf-auto-deploy relies heavily on Cloud Foundry's ability to deploy app manifests. Please refer to the Cloud Foundry Documentation for further information.
In order to put this docker image to work, simply provide mandatory environment variables and mount your directories for local artifacts, manifests as well as stage-related variables.
Example:
CF_CLIENT_ID="<secret>"
CF_CLIENT_SECRET="<secret>"
CF_TOKEN_URL="https://your.idp-server.com/oauth/token"
CF_CREDENTIALS_URL="https://your.oauth2-client.com/cf/credentials/url"
CF_SPACE="sandbox"
docker run --rm \
-v /local/path/to/artifacts:/tmp/artifacts \
-v /local/path/to/manifests:/tmp/manifests \
-v /local/path/to/stage-vars:/tmp/stage-vars \
-e CF_CLIENT_ID="${CF_CLIENT_ID}" \
-e CF_CLIENT_SECRET="${CF_CLIENT_SECRET}" \
-e CF_TOKEN_URL="${CF_TOKEN_URL}" \
-e CF_CREDENTIALS_URL="${CF_CREDENTIALS_URL}" \
-e CF_SPACE="${CF_SPACE}" \
dl0pes/cf-auto-deploy:1.0.3
The following directories must be mounted when a docker container is started from this docker image:
- /tmp/manifests: directory with all the manifests. If you have multiple manifest files you can set a number as prefix so the files are processed in the desired order.
- /tmp/stage-vars: directory with YAML files containing stage dependent variables. There must be a file that corresponds to the
CF_SPACEenvironment variable that you provide when starting the container. The file must be named{CF_SPACE}.yml. - /tmp/artifacts: directory with all deployment artifacts. These are the files that are referenced from your manifests.
You can mount a directory with shell scripts for pre- and post-processing. Pre-processing scripts need to start with the prefix pre- and post-processing scripts with the prefix post-.
Example:
docker run --rm \
-v /local/path/to/scripts:/tmp/scripts \
...
dl0pes/cf-auto-deploy:1.0.3
This command from above will automatically load your scripts and run them before and after the artifacts are deployed.
Please refer to the examples for two scripts which send slack notifications before and after the deployment process.
How to test your scripts?
Although cf-auto-deploy shall be used to rollout changes in a stadardized manner, sometimes it is simply needed to test your scripts.
Likewise arbitrary CF CLI commands could be run as one-off tasks. In both cases, you need to do stuff from the inside of the container.
cf-auto-deploy is prepared to handle such situations.
First run the docker container in interactive mode:
docker run --rm -it \
-v /local/path/to/artifacts:/tmp/artifacts \
-v /local/path/to/manifests:/tmp/manifests \
-v /local/path/to/stage-vars:/tmp/stage-vars \
-v /local/path/to/scripts:/tmp/scripts \
-e CF_CLIENT_ID="${CF_CLIENT_ID}" \
-e CF_CLIENT_SECRET="${CF_CLIENT_SECRET}" \
-e CF_TOKEN_URL="${CF_TOKEN_URL}" \
-e CF_CREDENTIALS_URL="${CF_CREDENTIALS_URL}" \
-e CF_SPACE="${CF_SPACE}" \
dl0pes/cf-auto-deploy:1.0.3 bash
When you are inside the docker container, you are connected to a bash and you can simply execute the initialization script:
% cf-init.sh
As a result you should see this output:
Authenticating...
OK
Targeted org <YOUR ORG>.
Targeted space <YOUR SPACE>.
API endpoint: https://your.api-server.com/v1/api
API version: 3.88.0
user: <generated>
org: <YOUR ORG>
space: <YOUR SPACE>
... and off you go
You can provide a comma-separated list of (task name):(task command) tupels.
Example:
docker run --rm \
...
-e APP_TASKS="sba-db-migrator:.java-buildpack/open_jdk_jre/bin/java de.dlopes.test.ExampleTask -k 256M -m 128M"
...
dl0pes/cf-auto-deploy:1.0.3
The task name needs to correspond with the name of your manifest file since these are matched based on their name. Once that tasks is matched to a manifest file, it will be automatically run after the artifact from the manifest has been successfully deployed.