aptos-node

Aptos blockchain node deployment

Homepage: https://aptoslabs.com/

Source Code

• https://github.com/aptos-labs/aptos-core

Values

Resource Descriptions

Below is a list of the Kubernetes resources created by this helm chart.

The resources created by this helm chart will be prefixed with the helm release name. Below, they are denoted by the <RELEASE_NAME> prefix.

StatefulSets:

• <RELEASE_NAME>-aptos-node-0-validator - The validator StatefulSet
• <RELEASE_NAME>-aptos-node-0-fullnode-e<ERA> - The fullnode StatefulSet

Deployments:

• <RELEASE_NAME>-aptos-node-0-validator - The HAProxy deployment

PersistentVolumeClaim:

• <RELEASE_NAME>-0-validator-e<ERA> - The validator PersistentVolumeClaim
• fn-<RELEASE_NAME>-0-fullnode-e<ERA>-0 - The fullnode PersistentVolumeClaim. Note the difference in naming scheme between valdiator and fullnode PVC, which is due to the fact that you can spin up multiple fullnodes, but only a single validator.

Services:

• <RELEASE_NAME>-aptos-node-0-validator-lb - Inbound load balancer service that routes to the validator
• <RELEASE_NAME>-aptos-node-0-fullnode-lb - Inbound load balancer service that routes to the fullnode

ConfigMaps:

• <RELEASE_NAME>-0 - The validator and fullnode NodeConfigs
• <RELEASE_NAME>-0-haproxy - The HAProxy configuration

NetworkPolicies:

• <RELEASE_NAME>-0-validator - The validator NetworkPolicy, which controls network access to the validator pods

ServiceAccounts:

• [optional] <RELEASE_NAME> - The default service account
• <RELEASE_NAME>-validator - The validator service account
• <RELEASE_NAME>-fullnode - The fullnode service account

[optional] PodSecurityPolicy:

• <RELEASE_NAME> - The default PodSecurityPolicy for validators and fullnodes
• <RELEASE_NAME>-haproxy - The PodSecurityPolicy for HAProxy

Common Operations

Check Pod Status

$ kubectl get pods

You should see at least 1/1 replicas running for the validator, fullnode, and HAProxy. If there are any restarts, you should see it in this view.

To see more details about a singular pod, you can describe it:

$ kubectl describe pod <POD_NAME>

Check the Pod Logs

$ kubectl logs <POD_NAME>

Check all services

$ kubectl get services

By default, the services are LoadBalancer type, which means that they will be accessible from the outside world. Depending on your kubernetes deployment/cloud, the public IP or DNS information will be displayed.

Scale Down Workloads

If you want to temporarily remove some of the workloads, you can scale them down.

# scale down the validator
kubectl scale statefulset <STS_NAME> --replicas=0

Advanced Options

Testnet Mode (Multiple Validators and Fullnodes)

For testing purposes, you may deploy multiple validators into the same cluster via .Values.numValidators. The naming convention is <RELEASE_NAME>-aptos-node-<INDEX>-validator, where <INDEX> is the index of the validator. Note that for each validator, you must provide genesis ConfigMaps for each, of the name: <RELEASE_NAME>-<INDEX>-genesis-e<ERA>. You may also deploy multiple fullnodes into the cluster via .Values.numFullnodeGroups and .Values.fullnode.groups. Each validator can have multiple fullnode groups, each with multiple replicas. The total number of fullnode groups can be limited via .Values.numFullnodeGroups.

Era

The .Values.chain.era is a number that is incremented every time the validator's storage is wiped. This is useful for testnets when the network is wiped.

Privileged Mode

For debugging purposes, it's sometimes useful to run the validator as root (privileged mode). This is enabled by .Values.enablePrivilegedMode.