The idea is to have a cluster, and try to simulate the possible problems that can happen in a cluster deployment: Latencies, network partition etc...
Make sure to have:
- Docker installed
- Docker compose installed
- Internet access (at least to download the needed images)
The project is structured as following:
.
│___README.md
│___docker-compose.yml
│
│
└───core1
│ │___ data
│ │___ backups
│ │___ conf
│ │___ import
│ │___ logs
│ │___ metrics
│
└───core2
│ │___ data
│ │___ backups
│ │___ conf
│ │___ import
│ │___ logs
│ │___ metrics
│
└───core3
│ │___ data
│ │___ backups
│ │___ conf
│ │___ import
│ │___ logs
│ │___ metrics
|
└───toxiproxy
│ │___ conf
| | |____ toxiproxy.json
│ │___ toxics
| | |____
│ │___ Dockerfile
docker-compose.yml: Will contain all the containers configuration
core{1,2,3}: Each directory stands for a Neo4j instance.
core{1,2,3}/data: All data-related content, such as databases, transactions, cluster-state (if applicable), dumps, and cypher script files (from the neo4j-admin restore command).
core{1,2,3}/backups: Directory where the backups will be stored using neo4j-admin.
core{1,2,3}/conf: The Neo4j configuration settings and the JMX access credentials.
core{1,2,3}/import: All the files that the command LOAD uses as sources to import data in Neo4j.
core{1,2,3}/logs: The Neo4j log files.
core{1,2,3}/metrics: The Neo4j built-in metrics for monitoring the Neo4j DBMS and each individual database.
toxiproxy: Directory for toxiproxy Dockerfile and configuration file.
A docker-compose file is ready with a deployment of a 3 cores cluster and a toxiproxy container.
Neo4j configuration files are located in ./core{1,2,3}/conf/neo4j.conf. You can fine tune them at your own needs (if you want to add a core, a replica or change the ports).
For ToxiProxy, there is a Dockerfile which adds a configuration file to the image when building it, and then start the toxiproxy server with this given configuration.
If you want to change the configuration, you can modify toxiproxy/conf/toxiproxy.json.
ToxiProxy will start on port 6666 to not conflict with the cluster configuration. Remeber to specify the correct port later for the CLI commands.
Before starting docker-compose, make sure that these directories are created and that the volumes points on them (the actual configuration will work if you don't change anything):
To start the cluster, simply run:
$> docker-compose -f docker-compose.yml up --build
To use the ToxiProxy cli, you need to be in the container:
$> docker exec -i -t toxiproxy sh
Once you are connected, go to:
$> cd /go/bin
You can execute different commands to activate/deactivate proxies, add/remove toxics, etc.
Please refer to toxiproxy/toxics folder to have all the commands to manage toxiproxy: