ScyllaDB Tablets DEMO

This project walks you through the following scenario:

Starting off with 3 ScyllaDB nodes
Spike in the number of requests
Add 3 additional ScyllaDB nodes to handle requests
Request volume goes down
Remove 3 nodes

Requirements

AWS account
Terraform
Ansible

Usage

Clone the repository

git clone https://github.com/zseta/tablets-scaling.git

Create EC2 instances in AWS (takes 4+ minutes)
```
terraform init
terraform plan
terraform apply
```
Open ScyllaDB Monitoring

One of the Terraform outputs is the link to Monitoring, similar to this:
```
Outputs:

monitoring_url = "http://<IP-ADDRESS>:3000"
[...]
```
There should be 1 active node and 5 unreachable nodes seen on the dashboard.
CD into the ansible folder to run playbooks
```
cd ansible
```
Start up a 3 node cluster
```
ansible-playbook 1_original_cluster.yml
```
Create schema and restore snapshot (takes 5+ minutes)
```
ansible-playbook 2_restore_snapshot.yml
```
Start cql-stress
```
ansible-playbook 3_stress.yml
```
Scale out (start up 3 more nodes)
```
ansible-playbook 4_scale_out.yml
```
Stop cql-stress
```
ansible-playbook 5_stop_stress.yml
```

Scale in (stop 3 nodes)

ansible-playbook ansible/6_scale_in.yml

Stop demo and remove AWS infrastructure
```
cd ..
terraform destroy
```

DEMO UI

You can also run the DEMO with a user interface. To access the UI, run Terraform as usual then open index.html (the link to this file is also an output of TF).