refactor: use named volume instead of bind mount in quickstart

Volume is the preferred method over bind mount (https://docs.docker.com/storage/volumes/) for persistent container data.
This also eliminates the need for the ugly chmod hack for elasticsearch and hopefully fixes #1650

README.md

@@ -39,7 +39,7 @@ This repository contains the complete source code for both DataHub's frontend &
 3. Clone this repo and `cd` into the root directory of the cloned repository.
 4. Run the following command to download and run all Docker containers locally:
     ```
-    cd docker/quickstart && source ./quickstart.sh
+    ./docker/quickstart/quickstart.sh
     ```
     This step takes a while to run the first time, and it may be difficult to tell if DataHub is fully up and running from the combined log. Please use [this guide](https://github.com/linkedin/datahub/blob/master/docs/debugging.md#how-can-i-confirm-if-all-docker-containers-are-running-as-expected-after-a-quickstart) to verify that each container is running correctly.
 5. At this point, you should be able to start DataHub by opening [http://localhost:9001](http://localhost:9001) in your browser. You can sign in using `datahub` as both username and password. However, you'll notice that no data has been ingested yet.

docker/quickstart/README.md

@@ -1,11 +1,9 @@
 # DataHub Quickstart
-To start all Docker containers at once, please run below command:
+To start all Docker containers at once, please run below command from project root directory:
 ```bash
-cd docker/quickstart && source ./quickstart.sh
+./docker/quickstart/quickstart.sh
 ```
 
-By default, data will be stored at `/tmp/datahub`, however it can be overwritten by specifying the DATA_STORAGE_FOLDER env var.
-
 At this point, all containers are ready and DataHub can be considered up and running. Check specific containers guide
 for details:
 * [Elasticsearch & Kibana](../elasticsearch)

docker/quickstart/docker-compose.yml

@@ -16,7 +16,7 @@ services:
       - "3306:3306"
     volumes:
       - ../mysql/init.sql:/docker-entrypoint-initdb.d/init.sql
-      - ${DATA_STORAGE_FOLDER}/mysql:/var/lib/mysql
+      - mysqldata:/var/lib/mysql
 
   zookeeper:
     image: confluentinc/cp-zookeeper:5.4.0
@@ -28,7 +28,7 @@ services:
       ZOOKEEPER_CLIENT_PORT: 2181
       ZOOKEEPER_TICK_TIME: 2000
     volumes:
-      - ${DATA_STORAGE_FOLDER}/zookeeper:/var/opt/zookeeper
+      - zkdata:/var/opt/zookeeper
 
   broker:
     image: confluentinc/cp-kafka:5.4.0
@@ -137,7 +137,7 @@ services:
       - xpack.security.enabled=false
       - "ES_JAVA_OPTS=-Xms1g -Xmx1g"
     volumes:
-      - ${DATA_STORAGE_FOLDER}/elasticsearch:/usr/share/elasticsearch/data
+      - esdata:/usr/share/elasticsearch/data
 
   kibana:
     image: docker.elastic.co/kibana/kibana:5.6.8
@@ -161,7 +161,7 @@ services:
       - "7474:7474"
       - "7687:7687"
     volumes:
-      - ${DATA_STORAGE_FOLDER}/neo4j:/data
+      - neo4jdata:/data
 
   # This "container" is a workaround to pre-create search indices
   elasticsearch-setup:
@@ -261,3 +261,9 @@ services:
 networks:
   default:
     name: datahub_network
+
+volumes:
+  mysqldata:
+  esdata:
+  neo4jdata:
+  zkdata:

docker/quickstart/quickstart.sh

@@ -1,10 +1,4 @@
 #!/bin/bash
 
-export DATA_STORAGE_FOLDER=${DATA_STORAGE_FOLDER:=/tmp/datahub}
-mkdir -p ${DATA_STORAGE_FOLDER}
-
-# https://discuss.elastic.co/t/elastic-elasticsearch-docker-not-assigning-permissions-to-data-directory-on-run/65812/4
-mkdir -p ${DATA_STORAGE_FOLDER}/elasticsearch
-sudo chmod 777 ${DATA_STORAGE_FOLDER}/elasticsearch
-
-docker-compose pull && docker-compose up --build
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+cd $DIR && docker-compose pull && docker-compose -p datahub up --build

docs/debugging.md

@@ -179,9 +179,9 @@ More discussions on the same issue https://github.com/docker/hub-feedback/issues
 ```
 docker rm -f $(docker ps -aq)
 ```
-2. Clear persistent storage for DataHub containers, assuming you didn't set `DATA_STORAGE_FOLDER` environment variable.
+2. Drop all DataHub's docker volumes.
 ```
-rm -rf /tmp/datahub
+docker volume rm -f $(docker volume ls -f name=datahub_*  -q)
 ```
 
 ## Seeing `Table 'datahub.metadata_aspect' doesn't exist` error when logging in