datayoga-io/datayoga

Adapt intergration-tests to run under docker

wheelly opened this issue · 3 comments

wheelly commented

There is a half-ready implementations of dev folder which must be accompinied by README.md file.

  1. Resources/connections.yml must point to the right container addresses
  2. Remove
    xxx_container = redis_utils.get_redis_oss_container(redis_utils.REDIS_PORT) xxx_container.start()
    And stop at the end of the test
  3. Think about automatically start/restart container (they are all stateless) upon test start - maybe omitted for now
  4. Add README.md file
wheelly commented

xxx_container = redis_utils.get_redis_oss_container(redis_utils.REDIS_PORT) xxx_container.start() - instead of removing maybe mock it with env variable not to do anything. Then the test will be intact. As to connections.yaml file - copy it to connections_docker.yaml and change all addresses there to point to the docker compose network domains.
The goal here is to be able after merging it to main to use it both in usual env (linux, WML) and in docker env

wheelly commented

@zalmane
# validate the connections connections_file = path.join(directory, "connections.yaml")
Can we add option for this file to be user defined? Like docker compose -f my_named_file.yaml

wheelly commented

There is no time to mess with docker now since it appears there are drivers compiled for arm in wheel (pip)