Here’s a really simple solution we found for zero downtime, blue–green style deploys with just docker-compose
and nginx
: https://engineering.tines.com/blog/simple-zero-downtime-deploys.
Turns out, a well-placed bash script is all you need here! This method saved us adding other dependencies to our stack. Here's the central piece:
reload_nginx() {
docker exec nginx /usr/sbin/nginx -s reload
}
zero_downtime_deploy() {
service_name=tines-app
old_container_id=$(docker ps -f name=$service_name -q | tail -n1)
# bring a new container online, running new code
# (nginx continues routing to the old container only)
docker-compose up -d --no-deps --scale $service_name=2 --no-recreate $service_name
# wait for new container to be available
new_container_id=$(docker ps -f name=$service_name -q | head -n1)
new_container_ip=$(docker inspect -f '{{range.NetworkSettings.Networks}}{{.IPAddress}}{{end}}' $new_container_id)
curl --silent --include --retry-connrefused --retry 30 --retry-delay 1 --fail http://$new_container_ip:3000/ || exit 1
# start routing requests to the new container (as well as the old)
reload_nginx
# take the old container offline
docker stop $old_container_id
docker rm $old_container_id
docker-compose up -d --no-deps --scale $service_name=1 --no-recreate $service_name
# stop routing requests to the old container
reload_nginx
}
Hope this can help others.