gem-hunterThis project shows how easy it is to connect an application in Kubernetes Engine to a Cloud SQL instance, using the Cloud SQL Proxy container as a sidecar container. You will deploy a Kubernetes Engine Cluster and a Cloud SQL Postgres instance, and use the Cloud SQL Proxy container to allow communication between them.
This demo shows how easy it is to connect an application in Kubernetes Engine to a Cloud SQL instance using the Cloud SQL Proxy container as a sidecar container. You will deploy a Kubernetes Engine (Kubernetes Engine) cluster and a Cloud SQL Postgres instance and use the Cloud SQL Proxy container to allow communication between them.
While this demo is focused on connecting to a Cloud SQL instance with a Cloud SQL Proxy container, the concepts are the same for any GCP managed service that requires API access.
The key takeaways for this demo are:
By default all Kubernetes Engine nodes are assigned the default Compute Engine service account. This service account is fairly high privilege and has access to many GCP services. Because of the way the Google Cloud SDK is setup, software that you write will use the credentials assigned to the compute engine instance on which it is running. Since we don't want all of our containers to have the privileges that the default Compute Engine service account has, we need to make a least-privilege service account for our Kubernetes Engine nodes and then create more specific (but still least-privilege) service accounts for our containers.
The only two ways to get service account credentials are 1.) through your host instance, which as we discussed we don't want, or 2.) through a credentials file. This demo will show you how to get this credentials file into your container running in Kubernetes Engine so your application has the privileges it needs.
The Cloud SQL Proxy allows you to offload the burden of creating and maintaining a connection to your Cloud SQL instance to the Cloud SQL Proxy process. Doing this allows your application to be unaware of the connection details and simplifies your secret management. The Cloud SQL Proxy comes pre-packaged by Google as a Docker container that you can run alongside your application container in the same Kubernetes Engine pod.
The application and its sidecar container are deployed in a single Kubernetes (k8s) pod running on the only node in the Kubernetes Engine cluster. The application communicates with the Cloud SQL instance via the Cloud SQL Proxy process listening on localhost.
The k8s manifest builds a single-replica Deployment object with two containers, pgAdmin and Cloud SQL Proxy. There are two secrets installed into the Kubernetes Engine cluster: the Cloud SQL instance connection information and a service account key credentials file, both used by the Cloud SQL Proxy containers Cloud SQL API calls.
The application doesn't have to know anything about how to connect to Cloud SQL, nor does it have to have any exposure to its API. The Cloud SQL Proxy process takes care of that for the application. It's important to note that the Cloud SQL Proxy container is running as a 'sidecar' container in the pod.
A Google Cloud account and project is required for this. Access to an existing Google Cloud project with the Kubernetes Engine service enabled If you do not have a Google Cloud account please signup for a free trial here.
Click the button below to run the demo in a Google Cloud Shell.
All the tools for the demo are installed. When using Cloud Shell execute the following command in order to setup gcloud cli. When executing this command please setup your region and zone.
gcloud initThis project will run on macOS, Linux, or in a Google Cloud Shell.
When not using Cloud Shell, the following tools are required.
gcloud config set project <your_p roject>The Google Cloud SDK is used to interact with your GCP resources. Installation instructions for multiple platforms are available online.
The kubectl CLI is used to interteract with both Kubernetes Engine and kubernetes in general. Installation instructions for multiple platforms are available online.
Prior to running this demo, ensure you have authenticated your gcloud client by running the following command:
gcloud auth application-default loginIf you don't have a Google Cloud account you can sign up for a free account.
Deployment is fully automated. In order to deploy you need to run create.sh. The script takes the following parameters, in order:
The script requires the following environment variables to be defined:
Here is what it looks like to run create.sh:
USER_PASSWORD=<password> PG_ADMIN_CONSOLE_PASSWORD=<password> ./create.sh <DATABASE_USER_NAME> <PGADMIN_USERNAME>
create.sh will run the following scripts:
Once create.sh is complete you need to run make expose to connect to
the running pgAdmin pod. make expose will port-forward to the running pod.
You can connect to the port-forwarded pgAdmin in your
browser. Use the <PGADMIN_USERNAME> in the "Email
Address" field and <PG_ADMIN_CONSOLE_PASSWORD> you defined earlier to login to the console.
From there you can click "Add New Server" and use the <DATABASE_USER_NAME> and
<USER_PASSWORD> you created earlier to connect to 127.0.0.1:5432.
Validation is fully automated. The validation script checks for the existence of the Cloud SQL instance, the Kubernetes Engine cluster, and the running pod. All of these resources should exist after the deployment script completes. Now that the application is deployed, we can validate these three deployments by executing:
make validateThe script takes the following parameters, in order:
A successful output will look like this:
Cloud SQL instance exists GKE cluster exists pgAdmin4 Deployment object exists
Teardown is fully automated. The teardown script deletes every resource created in the deployment script. In order to teardown you need to run,
make teardownIt will run teardown.sh which will destroy all of the resources created for this demonstration.
The script takes the following parameters, in order:
teardown.sh will run the following scripts:
When creating a Cloud SQL instance you get the error:
is the subject of a conflict: The instance or operation is not
in an appropriate state to handle the request.
You cannot reuse an instance name for up to a week after you have deleted an instance. https://cloud.google.com/sql/docs/mysql/delete-instance
This is not an officially supported Google product