I was following this guide: Spin Up a SingleStore Cluster on Docker Desktop in 10 Minutes to create a multi-node cluster and I realized that it only shows how to create one node using the memsql/cluster-in-a-box image.
If I create multiple containers using this image how can I make these containers aware of eachother so that they behave like a multi-node cluster? Is there any documentation for this process?
The best way to create a multi-node cluster is to use SingleStoreDB toolbox (especially sdb-deploy and sdb-admin) to gang together several logical machines into one cluster. See here for how to install toolbox:
The logical machines can be running on bare metal, VMs, or containers – your choice. You’ll have to set those up separately.
General SingleStoreDB installation instructions are here:
What do you mean by “logical machines” in this instance?
My goal is to create an offline virtual network on docker with 2 aggregator nodes and 2 leaf nodes where each node is running on separate containers on the same machine.
In this documentation post CLI Online Deployment - Red Hat Distribution · SingleStore Documentation
it says that you need to call the following command from the master-host to create the other nodes. I am assuming that this command uses ssh to remote access the machines and creates the nodes on the respective machines:
Hello, input from our support team. Cluster-in-a-box should be used as an image of an aggregator and a leaf node for dev/test proposes and that’s about it. It is not meant to be used as a base for a cluster running multiple nodes. To create a cluster running multiple nodes on one host using a containerized solution, you should use kubernetes (something like minikube should work fine).
I was able to use Ubuntu multipass to create 3 virtual server and then connect them to one another. Docker should work too. I have not played with Kubernetes yet.