Genestack Uplaoder

A HTTP server providing an API and a frontend for easy uploading to Genestack

Build & Run Docker Image -- by Fei

1. If any change is applied, version number (config.py) should be updated.

VERSION = "2.4"

2. hgi-docker.sh can build dev and production images automatically with right frontend/.env files.

It creates mercury/genestack-uploader:2.4.dev and mercury/genestack-uploader:2.4.prod with the appropriate configuration in the image.

sudo ./hgi-docker.sh 2.4

Check built imgages

sudo docker images
REPOSITORY                   TAG        IMAGE ID       CREATED          SIZE
mercury/genestack-uploader   2.4.prod   ff7989cfc80c   14 minutes ago   1.3GB
mercury/genestack-uploader   2.4.dev    2cdf3ab5dbe3   14 minutes ago   1.3GB

3. push the image to DockerHub

Log in docker with mercury first. (check secrets). Then push.

sudo docker login
sudo docker push mercury/genestack-uploader:2.4.dev
sudo docker push mercury/genestack-uploader:2.4.prod

4. log in dev and production (details on Confluence Swarm Page)

5. Update running image

docker pull mercury/genestack-uploader:2.4.dev
docker service scale mystack_genestack-uploader=1
docker service scale mystack_genestack-uploader=0
docker service update --image=mercury/genestack-uploader:2.4.dev mystack_genestack-uploader 
docker service scale mystack_genestack-uploader=1

Running with Docker 🐳 -- by Michael

1. Update frontend/.env and config.py if needed. (See docs/spec/main-spec-updates/1.md for discussion about base URL paths.)

2. Build the image, i.e.

docker build -t mercury/genestack-uploader:0.1.dev .

The tagging scheme used on this project is mercury/genestack-uploader:X.Y.{prod|dev}, where X.Y is the version number, and the final tag is prod or dev, describing which URL is specified in .env and config.py. (This is also described in more detail in docs/spec/main-spec-updates/1.md.)

If deploying in the HGI environments, the script hgi-docker.sh, passing the argument of the verison number will build the images as we need, for example ./hgi-docker.sh 2.0 will create mercury/genestack-uploader:2.0.dev and mercury/genestack-uploader:2.0.prod with the appropriate configuration in the image.

3. Run the image. You must:
   • provide the env variable GSSERVER with a value of either qc or default
   • link to config files /root/.genestack.cfg and /root/.s3cfg
   • link SSH key files to /root/.ssh/id_rsa_genestack and /root/.ssh/id_rsa_genestack.pub

Optional Environment Variables: - JOB_EXPIRY_HOURS: defaults to 168 (hours in a week) - this is how long a job should be kept after it has completed for it to be accessed using the /jobs/{uuid} API endpoint - LOG_LEVEL: one of DEBUG, INFO, WARNING, ERROR, CRITICAL (defaults to INFO) - the minimum level of logs to be reported

The app runs on port 5000 on a Docker network, so that can be used to forward it, such as in a nginx container.

To test, you can also expose port 5000, i.e.

docker run -p 80:5000 -e GSSERVER=default -v /home/ubuntu/genestack-uploader/configs:/root -d --name genestack-uploader mercury/genestack-uploader:0.1.dev

Version Numbering -- by Michael

There are two important version numbers to keep track of.

• The version number of the uploadtogenestack package
  • This is updated in the package's setup.py file, and is given a tag on GitLab
  • This tag number is then used in this repositories Dockerfile to install the right version
• The version number of this web app
  • This is updated in config.py and is also reflected in the Docker image tag

Ensure that these are kept up to date, as it really helps find bugs when you know which versions of each a particular image is running. The version numbers are displayed at the bottom of the homepage.