Migration commands need to be run before starting image, api and auth containers

[ToshKoevoets]

I've tried adding migrate command in deployment files, but I'm not sure I did it the right way, don't seem to get it working.

I've tried, on k8s/deployment/api.yaml

command: ['node migrate.js', 'npm start']

For dev environment I've solved it for now by adding command to npm start. Not sure what is better.

For api it is done by

node migrate.js

For image and auth it is managed with

knex migrate:latest

[diederikgithub]

@ToshKoevoets : The migrations are already done during container initialization with the initContainers command. So I think there should be no need for this?

[ToshKoevoets]

Migrations need to be run on every update of a container. This doesnt seem to happen now. Api errors unless I add Migration script. Init-db is only on creating right? I don't know what the correct way of doing this, best way to update a container would be one that has an order. What is proposed way of rolling out migrations with helm? Currenlty im autodeploying the development server via travis with a sh script like this: #!/bin/bash echo "$DOCKER_PASSWORD" | docker login -u "$DOCKER_USERNAME" --password-stdin; docker push ${DOCKER_PUBLIC_USERNAME}/${DOCKER_IMAGE_NAME}:${TRAVIS_BUILD_ID} docker tag ${DOCKER_PUBLIC_USERNAME}/${DOCKER_IMAGE_NAME}:${TRAVIS_BUILD_ID} ${DOCKER_PUBLIC_USERNAME}/${DOCKER_IMAGE_NAME}:devel #for now tag with development in future with specific bracch tags docker push ${DOCKER_PUBLIC_USERNAME}/${DOCKER_IMAGE_NAME}:devel curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/linux/amd64/kubectl chmod +x ./kubectl sudo mv ./kubectl /usr/local/bin/kubectl mkdir ${HOME}/.kube echo "$KUBE_CONFIG" | base64 --decode > ${HOME}/.kube/config echo "Deploying: deployment/${K8S_DEPLOYMENT_NAME}" kubectl set image deployment/${K8S_DEPLOYMENT_NAME} ${K8S_DEPLOYMENT_NAME}=${DOCKER_PUBLIC_USERNAME}/${DOCKER_IMAGE_NAME}:${TRAVIS_BUILD_ID} --record -n ${K8S_NAMESPACE} echo "Deployed: " ${DOCKER_PUBLIC_USERNAME}/${DOCKER_IMAGE_NAME}:${TRAVIS_BUILD_ID}

…

On Mon, 22 Jun 2020, 16:44 diederikgithub, ***@***.***> wrote: @ToshKoevoets <https://github.com/ToshKoevoets> : The migrations are already done during container initialization with the initContainers command. So I think there should be no need for this? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/Amsterdam/openstad-kubernetes/issues/29#issuecomment-647563705>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJEFAFQQXS72TVDZHTZX7LRX5U47ANCNFSM4ODYV63Q> .

[diederikgithub]

Updating the image must trigger the initContainers, so I'm not sure what is wrong. Tried it myself with:
kubectl set image deployment/openstad-api openstad-api=openstad/api:latest --record
and the logs say:
+ openstad-api-7447dd6c77-rzk5m › init-db-ready
openstad-api-7447dd6c77-rzk5m init-db-ready No new migrations

The change course is recorded (kubectl get deployment openstad-api -o yaml):
kubernetes.io/change-cause: kubectl set image deployment/openstad-api openstad-api=openstad/api:latest --record=true
Could you check that with your deployment?

[ToshKoevoets]

Okay will have to debug then, deploy with travis only started working when I add node migrate.js to the npm start command.

[diederikgithub]

@ToshKoevoets Did you manage to debug it?