1. docker build -t mkdn .
2. docker run --name my_first_docker -i -t mkdn
3. docker rm (container id can be found by trying to re-run the docker run command)
4. Log in to docker with docker login
5. Tag image on dockerhub with docker tag mkdn ebensma/mkdn:latest
6. Push image to dockerhub with docker push ebensma/mkdn:latest
7. Run with nextflow kuberun ebenz99/Kube-Test -latest
8. Save data to ./work directory with ./kube-runner/kube-save.sh deepgtex-prp .
9. nextflow kuberun login -v deepgtex-prp

Also good to know:

• delete unused docker images docker rmi -f $(docker images | grep "<none>" | awk "{print \$3}")
• run a deployment or pod kubectl apply -f workpod.yaml
• Log in to the NRP with nextflow kuberun login -v deepgtex-prp
• save pod data with kubectl cp saver-pod:a.txt .
• ssh into pod with kubectl exec -it <podname> -- /bin/bash
• copy data to pvc ./kube-runner/kube-load.sh task-pv-claim .
• delete helm stuff helm ls --all --short | xargs -L1 helm delete --purge
• update deployment helm upgrade a1 ./mychart --recreate-pods
• because ls isn't a bash command, it needs be before a /bin/bash entrypoint
• Minio Slate info: https://portal.slateci.io/applications/incubator/minio
• RUN vs CMD in Dockerfile https://stackoverflow.com/questions/37461868/difference-between-run-and-cmd-in-a-docker-file
• minio scripts: https://github.com/minio/minio/blob/master/docs/orchestration/kubernetes/
• PV(C) won't delete? Use this command to make it: kubectl patch pvc db-pv-claim -p '{"metadata":{"finalizers":null}}' (also check for volume attachments using kubectl get volumeattachments)

Deploying the Nextflow API works via pulling an image of the API from Dockerhub. To edit the api, fork the scidas/nextflow-api repo, edit its Dockerfile, and make sure that the values.yaml of the Helm charts is targeting the image you've created.

• Peruse the incubator - >slate app list --dev
• set up minio configuration -> slate app get-conf minio --dev -o minio.yaml
• create a secret -> slate secret create ekey --group <group> --cluster <cluster> --from-literal=accesskey=<username> --from-literal=secretkey=<secret>
• edit the yaml to use service type NodePort and not ClusterIP. Give it an instance name. Also give it the name of the secret you've created
• installing minio -> slate app install minio --cluster <YOUR CLUSTER> --group <YOUR GROUP> --conf minio.yaml --dev

• docker run -p 9000:9000 --user $(id -u):$(id -g) --name minio1 -e "MINIO_ACCESS_KEY=<key>" -e "MINIO_SECRET_KEY=<key2>" -v ${HOME}/data:/data minio/minio server /data

AWS scope in Nextflow useless for writing nextflow logs to Minio PVC; will have to use -with-weblog option to POST to PVC (possible including keys in query string)