idabmat / quantum-nodes

Running Quantum on multiple nodes

This example is taken from quantum_swarm but uses a single mix application instead of an Phoenix umbrella application.

Setup

Just run docker-compose build to set everything up.

Libraries

• Quantum for managing CRON tasks.
• Distillery to generate the application which runs on the container to simulate a production environment.
• libcluster for automatic cluster formation. Different strategies can be used to achieve it. In this case we poll from the DNS A records to find other availble nodes to connect to.

Usage

Start the application with docker-compose up. After a while, you should start to receives messages like the following every second:

sample_app_1  | 21:10:40.001 [warn]  Hello from [email protected]

Then in a separate session, you can issue docker-compose up --scale sample_app=X -d where X is the new total number of running instances.

sample_app_3  | 
sample_app_3  | 21:15:43.004 [warn]  Hello from [email protected]
sample_app_3  | 
sample_app_3  | 21:15:44.003 [warn]  Hello from [email protected]
sample_app_2  | 
sample_app_2  | 21:15:45.001 [warn]  Hello from [email protected]
sample_app_2  | 
sample_app_2  | 21:15:46.001 [warn]  Hello from [email protected]
sample_app_2  | 
sample_app_2  | 21:15:47.002 [warn]  Hello from [email protected]
sample_app_1  | 
sample_app_1  | 21:15:48.003 [warn]  Hello from [email protected]

Todo

A couple of things might need some investigating.

On Scale Up

When scaling up, I looks like the first instance receives a :shutdown. I still need to figure out if this comes either from the way docker-compose up --scale works or not. In either case this could be problematic if the shutdown happens before state synchronization between the nodes.

sample_app_1  | 21:11:38.278 [error] GenServer SampleApp.Scheduler.ExecutorSupervisor terminating
sample_app_1  | ** (stop) no process: the process is not alive or there's no process currently associated with the given name, possibly because its application isn't started
sample_app_1  | Last message: {:DOWN, #Reference<0.1320933371.735313921.234648>, :process, #PID<0.1278.0>, :noproc}
sample_app_1  |
sample_app_1  | 21:11:38.278 [warn]  [swarm on [email protected]] [tracker:handle_topology_change] handoff failed for SampleApp.Scheduler.ExecutionBroadcaster: {:shutdown, {GenServer, :call, [#PID<0.1278.0>, {:swarm, :begin_handoff}, 5000]}}

On scale down

Any remaining node is considered the source of truth. We probably should investigate if there is the possibility of losing information in such a case.