Docker Enterprise 3.0: Now with New Built-in Docker cluster CLI Plugin

Estimated Reading Time: 8 minutes

Last Dockercon, dozens of new Docker CLI Plugin were introduced. All of these CLI plugins will be available in upcoming Docker Enterprise 3.0 GA release this year. Docker Desktop Enterprise 3.0 Public Beta was made available soon after Dockercon event during 2nd week of May 2019. This public beta consists of Desktop Enterprise 2.0.0.4-ent, Universal Control Plane 3.2, Docker Trusted Registry 2.7, and Engine Enterprise 19.03.0. Similar to previous deployments, Docker Enterprise components except Docker Engine are deployed as containers. Please note that only a limited subset of operating systems have been tested for the current beta release, including RHEL 7.6, and Ubuntu 16.04 and 18.04, and Windows Server 2019.

What is DCI all about?

One of the primary focus of this public beta is enhancement around expanding choices. Docker Certified Infrastructure(DCI) is Docker’s prescriptive approach to deploying Docker Enterprise Edition on a range of infrastructures. DCI is designed to automate and reliably deliver a secure, enterprise-ready container platform, integrated with your existing management and infrastructure tools.

Is DCI targeted only for Enterprise customers?

The short answer is “Yes”. DCI is installed in Docker Engine – Enterprise and Desktop Enterprise by default. DCI provides a declarative way to build and manage Docker clusters. It implements a Docker CLI plugin that exposes a `docker cluster` top-level command, and lets you define a cluster in a YAML file.

How does it work?

At a high-level, you define a cluster in a YAML file and instantiate it with `docker cluster create`. The DCI back-end then performs the hard work of building the cluster.

What Platform does it support?

DCI currently supports building and managing clusters on AWS during the Public beta with upcoming support Azure, and VMware vSphere by General Availability.

In my last blog, I talked about “What’s New in Docker Desktop Enterprise 3.0” which introduced a new way to build, share and run multi-service apps on any infrastructure with Docker Applications. Under this blog post, I will showcase how to get started with docker clusterCLI plugin

Pre-requisite:

[Captains-Bay]🚩 >  aws --version
aws-cli/1.11.107 Python/2.7.10 Darwin/17.7.0 botocore/1.5.70
[Captains-Bay]🚩 >  
  • AWS Access Keys

If you already have an `~/.aws/credentials` file, you can skip this step.  Use the `aws configure` command to specify your AWS credentials.

You will require a Docker ID with access to a Docker UCP subscription either:

  • Docker Enterprise 3.0 Beta License for Docker Enterprise 3.0 Beta
  • An active Docker Enterprise license (paid or trial) to install generally available Docker Enterprise version

Also, An AWS account with security credentials you will need AWS credentials with the following IAM policies:

  • AmazonEC2FullAccess
  • AmazonElasticFileSystemFullAccess
  • AmazonRoute53DomainsFullAccess
  • AmazonS3FullAccess
  • IAMFullAccess (for creating instance profiles with roles and policies)

  • Under Docker Beta registration page, sign in with your DockerID
  • Once you complete your registration, you will see the links for Docker Desktop Enterprise for Mac and Windows. Download your preferred software based on your desktop OS.

Installing Docker Desktop Enterprise

You can directly download Desktop Enterprise for Mac too with the below link:

https://download.docker.com/mac/enterprise/Docker.pkg

To install double click the .pkg file. For Mac administrators, the following command line options support fine tuning and mass installation, after which Docker Desktop Enterprise can be run from the Applications folder on each individual machine.

sudo installer -pkg Docker.pkg -target /

The license file must then be either installed in the following location:

~/Library/Group Containers/group.com.docker/docker_subscription.lic


Or can be provided in the UI when starting the application for the first time (/Applications/Docker.app).

Click on “whale” icon which appear at the right top of the screen to verify if Docker Desktop Enterprise comes up well.

[Captains-Bay]🚩 >  docker version
Client: Docker Engine - Enterprise
 Version:           19.03.0-beta4
 API version:       1.40
 Go version:        go1.12.5
 Git commit:        d9934ea
 Built:             Tue May 14 06:40:00 2019
 OS/Arch:           darwin/amd64
 Experimental:      false

Server: Docker Engine - Enterprise
 Engine:
  Version:          19.03.0-beta4
  API version:      1.40 (minimum version 1.12)
  Go version:       go1.12.5
  Git commit:       d9934ea
  Built:            Tue May 14 06:46:25 2019
  OS/Arch:          linux/amd64
  Experimental:     true
 containerd:
  Version:          v1.2.6
  GitCommit:        894b81a4b802e4eb2a91d1ce216b8817763c29fb
 runc:
  Version:          1.0.0-rc8
  GitCommit:        425e105d5a03fabd737a126ad93d62a9eeede87f
 docker-init:
  Version:          0.18.0
  GitCommit:        fec3683

Login to Docker Hub

Login to Docker Hub with a Docker ID that has access to a Docker EE/UCP repository.

[Captains-Bay]🚩 >  docker login
Login with your Docker ID to push and pull images from Docker Hub. If you don't have a Docker ID, head over to https://hub.docker.com to create one.
Username: ajeetraina
Password: 
Login Succeeded

Testing the inbuilt docker cluster CLI Plugin

[Captains-Bay]🚩 >  docker cluster version
Version:  v0.3.0
Commit:   dc3d07a
Build:    Plugin

Cluster Declaration

It’s time to declare our cluster. We’ll use the following YAML file to deploy a new cluster to AWS. By default, `docker cluster create` will look for a cluster.yml file in the current working directory.  Alternatively, you can give the file any name you choose. Let’s create a cluster.yml file with the following contents of a simple cluster definition.  The below YAML will allow you to install Docker Enterprise 3.0 beta on 1 manager and 1 DTR node.

variable:
  region: us-east-1
  subscription_url: https://storebits.docker.com/ee/m/sub-zxxxx/  ## Don't forget to add / at the end as shown 
  ucp_password:
    type: "prompt"

provider:
  aws:
    region: ${region}

cluster:
  engine:
    url: ${subscription_url}
    version: "ee-test-19.03"
  ucp:
    version: "docker/ucp:3.2.0-beta4"
    username: "admin"
    password: ${ucp_password}
  dtr:
    version: "docker/dtr:2.7.0-beta4"
 
resource:
  aws_instance:
    managers:
      quantity: 1
    registry:
      quantity: 1

Let us go through each of the below section one by one –

The YAML has four top-level resources:

– variable- provider
– cluster
– resource

The `variable` section declares variables that will be used in the cluster declaration.  The ucp_password uses type “prompt” to indicate that `docker cluster` will request a value at cluster creation.

The `provider` section declares that this cluster will be deployed in AWS, and references the region parameter.

The `cluster` section defines the Docker Engine and UCP versions to deploy. It also specifies the UCP admin credentials to apply to the cluster.

The `resource` section requests a single AWS instance to be configured as a UCP manager.

Spinning up Docker Enterprise 3.0 on AWS Platform

[Captains-Bay]🚩 >  docker cluster create -f cluster.yml --log-level debug
Please provide a value for ucp_password
DEBU[0009] Image Ref: sha256:ea8a7a832f839d48f478e37602cb7f67207be6f612c3a00aeafa42ca9f155214 
DEBU[0009] Generating public/private rsa key pair.      
DEBU[0010] Your identification has been saved in /data/keys/ssh/id_rsa. 
DEBU[0010] Your public key has been saved in /data/keys/ssh/id_rsa.pub. 
DEBU[0010] The key fingerprint is:                      
DEBU[0010] SHA256:CnQ4M5/f+2AOXj+azUVReBXXXXX cluster@a1f8091cbb6a 
DEBU[0010] The key's randomart image is:                
DEBU[0010] +---[RSA 2048]----+                          
DEBU[0010] |       ..      +o|                          
DEBU[0010] |     .  ..    o.+|                          
DEBU[0010] |    * .o.     .o.|                          
DEBU[0010] |   . *+.oo    o. |                          
DEBU[0010] |    ..o=S    ... |                          
DEBU[0010] |     .oo o  o.   |                          
DEBU[0010] |      ..+.Oo  .  |                          
DEBU[0010] |      o+.E.B..   |                          
DEBU[0010] |     o+oo =o=.   |                          
DEBU[0010] +----[SHA256]-----+                          
DEBU[0010] Planning cluster on aws       

Sit back & Relax ! This is going to take couple of minutes to bring up your Docker Enterprise 3.0

Troubleshooting Tips:

In case you encounter issue around unable to pull dockereng/cluster:v0.3.0

there is a quick workaround. Reason – The dockereng/cluster:v0.3.0 is a private Docker image which would fail to get pulled from Dockerhub. You might need to follow the below steps:

[Captains-Bay]🚩 >  docker pull docker/cluster:v0.3.0
v0.3.0: Pulling from docker/cluster
bdf0201b3a05: Pull complete 
227965e0be77: Pull complete 
656c27da0276: Downloading  10.18MB/98.87MB
6bc49ae6e7fa: Download complete 
ddbd7883b3bf: Download complete 
90dd03face76: Download complete 
cb5cae322035: Download complete 
c0c9485136e8: Download complete 
a5ab55def61b: Download complete 
ddbd7b624dc0: Download complete

Now you need to tag it to dickering/cluster:v0.3.0so as to let CLI plugin consider it locally and pick it up for building the cluster.

docker tag docker/cluster:v0.3.0 dockereng/cluster:v0.3.0

Please note that this issue has been fixed under cluster CLI version 0.3.3. By now, you should be able to see the below window while accessing it over the browser.


Once you upload License, you should be able to access Docker Enterprise 3.0 UI as shown below:

Inspecting the cluster

You can use docker cluster ls to list out the cluster. Even you can inspect the cluster using the below command:

[Captains-Bay]🚩 >  docker cluster inspect fervent_taussig
name: fervent_taussig
shortid: 67fb8cb05043
variable:
  region: us-east-1
  subscription_url: https://storebits.docker.com/ee/m/sub-a3dd83ed-d9db-440f-a175-e11347fb1037/
  ucp_password: Oracle9ias
provider:
  aws:
    region: us-east-1
    tags:
      pet: "true"
      project: CSG-DCI
    version: ~> 1.0
cluster:
  dtr:
    version: docker/dtr:2.7.0-beta4
  engine:
    storage_volume: /dev/xvdb
    url: https://storebits.docker.com/ee/m/sub-a3dd83ed-d9db-440f-a175-e11347fb1037/
    version: ee-test-19.03
  registry:
    url: https://index.docker.io/v1/
    username: ajeetraina
  ucp:
    username: admin
    version: docker/ucp:3.2.0-beta4
resource:
  aws_instance:
    managers:
      _running:
        managers_ids:
        - i-088036137bdf5564a
        managers_ips:
        - 35.170.33.58
      instance_type: t2.xlarge
      os: Ubuntu 16.04
      quantity: 1
      role: manager
    registry:
      _running:
        registry_ids:
        - i-016770ea989a55a0a
        registry_ips:
        - 18.208.208.51
      instance_type: t2.xlarge
      os: Ubuntu 16.04
      quantity: 1
      role: dtr

Using context switching to switch from Docker Desktop to remote AWS cluster

[Captains-Bay]🚩 >  docker context ls
NAME                DESCRIPTION                               DOCKER ENDPOINT               KUBERNETES ENDPOINT                ORCHESTRATOR
default *           Current DOCKER_HOST based configuration   unix:///var/run/docker.sock   https://localhost:6443 (default)   swarm
fervent_taussig     fervent_taussig                           tcp://35.170.33.58:443                                           
[Captains-Bay]🚩 >  docker context use fervent_taussig
fervent_taussig
Current context is now "fervent_taussig"
[Captains-Bay]🚩 >  docker node ls
ID                            HOSTNAME                       STATUS              AVAILABILITY        MANAGER STATUS      ENGINE VERSION
fbe5k12hyoit5qcdtatamz907 *   ip-172-31-9-19.ec2.internal    Ready               Active              Leader              19.03.0-beta4
hs8jz9vnuwqjjjukzh9s2rejc     ip-172-31-10-73.ec2.internal   Ready               Active                                  19.03.0-beta4
[Captains-Bay]🚩 >

As you can see, it shows up UCP nodes cluster running on remote AWS Cloud Platform.

Open up the browser and you shall be able to access Docker Enterprise v3.2.0-beta4 Release.

In my next blog post, I will talk around docker registry as well as docker gmsa CLI Plugin. Stay tuned !




Clap

(5)