docker-compose 3.4 support

[blaggacao]

The depends_on flag was conceptually simplified only representing "docker running"
Readiness testing was made responsibility of the application.

[blaggacao]

see: https://docs.docker.com/compose/startup-order/

#!/bin/bash
# wait-for-postgres.sh

set -e

host="$1"
shift
cmd="$@"

until psql -h "$host" -U "postgres" -c '\q'; do
  >&2 echo "Postgres is unavailable - sleeping"
  sleep 1
done

>&2 echo "Postgres is up - executing command"
exec $cmd

command: ["./wait-for-postgres.sh", "master", "other_command"]

[vinhbuivan]

Checkout with this docker-compose version. I tested and it run on docker stack.
docker-compose.zip

version: '3.7'
services:
  master:    
    image: 'citusdata/citus:8.3.2'
    ports: ["${MASTER_EXTERNAL_PORT:-5432}:5432"]
    labels: ['com.citusdata.role=Master', 'com.docker.compose.project=citus']
    healthcheck:
  worker:
    image: 'citusdata/citus:8.3.2'
    labels: ['com.citusdata.role=Worker', 'com.docker.compose.project=citus']
    depends_on: manager      
  manager:    
    image: 'citusdata/membership-manager:0.2.0'
    volumes: ['/var/run/docker.sock:/var/run/docker.sock']
	labels: ['com.citusdata.role=Manager', 'com.docker.compose.project=citus']
    depends_on: master      

[mubaidr]

Issue #124 Worker node not being added is caused by this. Because manager service tries to connect to master service before it is ready and fails. But setting it to restart: on-failure will make sure it tires multiple times and eventually it will work.

More info: #124 (comment)

[hanefi]

I created two PRs that aims to resolve this issue:

1. Start polling for connections to coordinator node membership-manager#9 implements a polling mechanism in Membership Manager, so that (a) it will detect the readiness of the coordinator node, (b) properly report that it is ready to accept new Citus worker services
2. Implement service dependencies for Docker-Compose V3 #187 (a) updates the docker images to properly detect dependencies, (b) introduce Compose V3 definitions

Once they are merged, this issue should be resolved