Canvas LMS - Local Deployment setup

Local deployment of Canvas LMS using Helm Charts

1. Install kubectl.

• Install minikube.
• Start the minikube cluster
  • minikube start
• Verify minikube cluster is running.
  • minikube status
• Install Helm
• In the terminal window, change directory to the canvas-lms/helm directory.
• Run the following command to install the Canvas LMS Helm Chart
  • helm install --generate-name ./canvas-lms-app
• Verify that all resources are deployed and running
  • kubectl get all
  • This should show 6 total services: 5x services above (eg, postgres-service.yaml) + kubernetes service
    • The 5x services listed above should be listed as 'pods' and next to them the status should show "Running"
    • On first run this could take 20 minutes or so for status to update from "ContainerCreating" to "Running" as large images are downloaded.
    • You can type kubectl describe pod to see the status
      • look for 'events' under each service and there should be a status such as: 'scheduled', 'pulling', 'pulled', 'created', 'started'
      • if the last event is 'pulling' then you know the image is downloading and you must wait.
• Perform initial Canvas setup
  • kubectl exec -it canvas-lms-0 -- sh
    • After executing this your terminal will be in the shell for Canvas
  • bundle exec rake db:create
  • bundle exec rake db:migrate
    • Wait for a while, after a few mins you should see several tables being created and migrations completed in stdout
    • This will take ~20min in total
    • NOTE: This sometimes fails on a migration, if it does just re-run bundle exec rake db:migrate to complete the migrations
  • bundle exec rake db:initial_setup
    • This will require you to answer some questions in the terminal (eg, admin username/password)
  • bundle exec rake canvas:compile_assets
    • This can take some time (~20min)
  • bundle exec rake brand_configs:generate_and_upload_all
    • This should be far quicker than previous steps
  • exit
• Get the URL associated with the canvas-lms-0 service and open in a browser.
  • minikube service list
• Verify login functionality
  • Attempt to login using the email address and password you supplied during the initial setup of canvas
  • NOTE: If you encounter an error during login run the following command to resolve the issue
    • kubectl rollout restart sts canvas-lms

Local deployment of Canvas LMS using Kubernetes and Minikube

1. Install kubectl.

• Install minikube.
• Start the minikube cluster
  • minikube start
• Verify minikube cluster is running.
  • minikube status
• In the terminal window, change directory to the canvas-lms directory.
• Run the following commands to create Canvas services.
  • kubectl apply -f postgres-configmap.yaml
  • kubectl apply -f postgres-secret.yaml
  • kubectl apply -f postgres-service.yaml
  • kubectl apply -f mail-configmap.yaml
  • kubectl apply -f mail-secret.yaml
  • kubectl apply -f mail-service.yaml
  • kubectl apply -f redis-service.yaml
  • kubectl apply -f canvas-configmap.yaml
  • kubectl apply -f canvas-service.yaml
  • kubectl apply -f canvas-worker-service.yaml
• Verify all Canvas services are running
  • kubectl get all
    • This should show 6 total services: 5x services above (eg, postgres-service.yaml) + kubernetes service
    • The 5x services listed above should be listed as 'pods' and next to them the status should show "Running"
    • On first run this could take 20 minutes or so for status to update from "ContainerCreating" to "Running" as large images are downloaded.
    • You can type kubectl describe pod to see the status
      • look for 'events' under each service and there should be a status such as: 'scheduled', 'pulling', 'pulled', 'created', 'started'
      • if the last event is 'pulling' then you know the image is downloading and you must wait.
• Perform initial Canvas setup
  • kubectl exec -it canvas-lms-0 -- sh
    • After executing this your terminal will be in the shell for Canvas
  • bundle exec rake db:create
  • bundle exec rake db:migrate
    • Wait for a while, after a few mins you should see several tables being created and migrations completed in stdout
    • This will take ~20min in total
    • This sometimes fails on a migration, if it does just re-run bundle exec rake db:migrate to complete the migrations
  • bundle exec rake db:initial_setup
    • This will require you to answer some questions in the terminal (eg, admin username/password)
  • bundle exec rake canvas:compile_assets
    • This can take some time (~20min)
    • you may need to kill this process if it gets stuck at the end after outputting "Done [###]s"
  • bundle exec rake brand_configs:generate_and_upload_all
    • This should be far quicker than previous steps
  • exit
• Get the URL associated with the canvas-lms-0 service and open in a browser.
  • minikube service list
  • NB: if you are using OSX, this final step will NOT work. Various links that may help in de-bugging/workarounds:
    • kubernetes/minikube#11193
    • https://stackoverflow.com/questions/66873437/access-minikube-cluster-on-the-same-network/66945182#66945182
    • https://minikube.sigs.k8s.io/docs/handbook/accessing/ (see DNS resolution heading re: docker driver not working on Macs)
    • https://kubernetes.io/docs/tasks/access-application-cluster/port-forward-access-application-cluster/
• Verify login functionality
  • Attempt to login using the email address and password you supplied during the initial setup of canvas
  • NOTE: If you encounter an error during login run the following command to resolve the issue
    • kubectl rollout restart sts canvas-lms

Local deployment of Canvas LMS using docker and docker-compose.

1. Setup docker for your environment.

2. Setup docker-compose for your environment.

3. In GitLab, create or use an existing, personal access token.

4. In a terminal window, docker login to the GitLab registry using your GitLab username and personal access token created in step 3. docker login registry.gitlab.com

5. Pull the Canvas LMS image from the GitLab education-module container registry. docker pull registry.gitlab.com/onestepprojects/education-module/canvas-lms:stable

6. Verify the Canvas LMS image is in your docker images repository. docker images
   You should see registry.gitlab.com/onestepprojects/education-module/canvas-lms listed as a repository, and stable listed as the tag.

7. In a terminal window, change directory to the location of the education-module, and verify the existence of the canvas-lms directory.

8. In the terminal window, change directory to the canvas-lms directory.

9. In the terminal window, spin up the images and performance initial setup commands:

   1. mkdir .data && sudo chmod -R 777 .data -- grant full access to the data store for the container
   2. docker-compose up -d db -- starts up all the database container identified in the docker-compose.yml file.
   3. docker-compose run --rm app bundle exec rake db:create db:initial_setup -- initialized the canvas lms database, this only needs to run once
   4. docker-compose down -- shutdown all containers
   5. sudo nano /etc/hosts and add the line 127.0.0.1 docker_canvas_app to the file, ctrl-x to exit
   6. docker-compose up -d -- starts up all the containers identified in the docker-compose.yml file.
   7. docker ps -a -- shows a list of running docker containers, confirm one of the running containers is named canvas-lm_app_1
   8. docker exec -u root -it canvas-lms_app_1 /bin/bash -- opens a shell into the canvas-lms_app_1 container
   9. bundle exec rake canvas:compile_assets -- creates all canvas assets
   10. bundle exec rake brand_configs:generate_and_upload_all -- generate css assets
   11. exit -- exist the contain shell

10. If all steps are completed successfully, open a browser and vist http://docker_canvas_app:8900/

Start Server

docker-compose up -d

Check Logs

app

docker logs -f docker-canvas_app_1

more detailed app logs

docker exec -it docker-canvas_app_1 tail -f log/production.log

worker

docker logs -f docker-canvas_worker_1

db

docker logs -f docker-canvas_db_1

redis

docker logs -f docker-canvas_redis_1

mail

docker logs -f docker-canvas_mail_1

Shutdown Server

docker-compose down