Skip to content

Quick Start

Build and Deploy Gloo

This quick start builds and deploys Gloo to minikube. You need Docker, Helm and kubectl

Step 1 Install

Please follow the install guide to install thetool.

Step 2 Initialize

Create a working directory for thetool and initialize the directory. Please pass your Docker user id.

mkdir thetool-gloo
cd thetool-gloo
thetool init -u axhixh

This should show the following message.

Initializing current directory...
Adding default repositories...
Added repository with commit hash 282a844ea3ed2527f5044408c9c98bc7ee027cd2

You can verify it has initialized correctly by getting the feature list.

thetool list

This should return a list of Gloo features that are enabled by default.

Name:             aws_lambda
Gloo Directory:   aws
Envoy Directory:  aws/envoy
Enabled:          true

Name:             google_functions
Gloo Directory:   google
Envoy Directory:  google/envoy
Enabled:          true

Name:             kubernetes
Gloo Directory:   kubernetes
Envoy Directory:  
Enabled:          true

Name:             transformation
Gloo Directory:   transformation
Envoy Directory:  transformation/envoy
Enabled:          true

Step 3 Build

You can build Gloo and its components by using the build command. For the quick start we are going to build all the components.

thetool build all

The build command builds and publishes newly created Docker images.

Building and publishing with 4 features
Building Envoy...
Publishing Envoy...
Pushed Envoy image axhixh/envoy:be0d5f72
Building Gloo...
Adding plugins to Gloo...
Constraining plugins to given revisions...
Publishing Gloo...
Pushed Gloo image axhixh/gloo:be0d5f72
Building gloo-function-discovery...
Publishing gloo-function-discovery...
Pushed gloo-function-discovery image axhixh/gloo-function-discovery:644fefd
Building gloo-ingress-controller...
Publishing gloo-ingress-controller...
Pushed gloo-ingress-controller image axhixh/gloo-ingress-controller:90f2b21
Building gloo-upstream-discovery...
Publishing gloo-upstream-discovery...
Pushed gloo-upstream-discovery image axhixh/gloo-upstream-discovery:12b4753

thetool builds Gloo and its components using Docker containers. The first build can take some time, specially on macOS. You can speed up the build by giving the Docker machine more memory and CPUs. Future build will use the cache and be a lot quicker.

You can also run build command in verbose mode to see the progress.

thetool build all -v
When building Envoy, Bazel build can fail with the error message: gcc: internal compiler error: Killed (program cc1plus), if the virtual machine is out of memory. You can fix it by either reducing the number of cores or increasing the RAM on Docker VM. You can set the VM to 2GB RAM and 2 CPUs for a working configuration.

Step 4 Deploy

You can use the deploy command to deploy Gloo to Kubernetes.

thetool deploy k8s

You should see the following output.

Downloading Gloo chart from
Building with 4 features
Generating Helm Chart values...

You can verify this using Helm and Kubectl.

helm list
NAME        REVISION    UPDATED                     STATUS      CHART       NAMESPACE  
goodly-duck 1           Thu Mar  8 13:38:21 2018    DEPLOYED    gloo-0.1.0  gloo-system
kubectl get pods -n gloo-system
NAME                                              READY     STATUS    RESTARTS   AGE
goodly-duck-function-discovery-84c67868bd-9hbtl   1/1       Running   0          1m
goodly-duck-gloo-5f7f975c75-754kf                 1/1       Running   0          1m
goodly-duck-ingress-6cfcbd4784-pbpzw              1/1       Running   0          1m
goodly-duck-ingress-controller-6fc9f58b74-h784s   1/1       Running   0          1m
goodly-duck-upstream-discovery-86dfcd79-vv79b      1/1       Running   0          1m