This guide is focused on deploying Scylla on EKS with improved performance. Performance tricks used by the script won't work with different machine tiers. It sets up the kubelets on EKS nodes to run with static cpu policy and uses local sdd disks in RAID0 for maximum performance.
Most of the commands used to setup the Scylla cluster are the same for all environments As such we have tried to keep them separate in the general guide.
If you don't want to run the commands step-by-step, you can just run a script that will set everything up for you:
# Edit according to your preference
EKS_ZONE=us-east-1a
# From inside the examples/eks folder
cd examples/eks
./eks.sh -z "$EKS_ZONE"
After you deploy, see how you can benchmark your cluster with cassandra-stress.
First of all, we export all the configuration options as environment variables. Edit according to your own environment.
EKS_ZONE=us-east-1a
CLUSTER_NAME=scylla-demo
For this guide, we'll create a EKS cluster with the following:
- A NodeGroup of 3
i3-2xlarge
Nodes, where the Scylla Pods will be deployed. These nodes will only accept pods havingscylla-clusters
toleration.
- name: scylla-pool
instanceType: i3.2xlarge
desiredCapacity: 3
labels:
pool: "scylla-pool"
taints:
role: "scylla-clusters:NoSchedule"
ssh:
allow: true
kubeletExtraConfig:
cpuManagerPolicy: static
- A NodeGroup of 4
c4.2xlarge
Nodes to deploycassandra-stress
later on. These nodes will only accept pods havingcassandra-stress
toleration.
- name: cassandra-stress-pool
instanceType: c4.2xlarge
desiredCapacity: 4
labels:
pool: "cassandra-stress-pool"
taints:
role: "cassandra-stress:NoSchedule"
ssh:
allow: true
- A NodeGroup of 1
i3.large
Node, where the monitoring stack and operator will be deployed.
- name: monitoring-pool
instanceType: i3.large
desiredCapacity: 1
labels:
pool: "monitoring-pool"
ssh:
allow: true
Script requires several dependencies:
- Helm - See: https://docs.helm.sh/using_helm/#installing-helm
- eksctl - See: https://docs.aws.amazon.com/eks/latest/userguide/getting-started-eksctl.html
- kubectl - See: https://kubernetes.io/docs/tasks/tools/install-kubectl/
We deploy the local volume provisioner, which will discover their mount points and make them available as PersistentVolumes.
helm install local-provisioner examples/common/provisioner
Now you can follow the generic guide to launch your Scylla cluster in a highly performant environment.
Once you are done with your experiments delete your cluster using following command:
eksctl delete cluster <cluster_name>