Module: PILOT

• 1: Module: DuckDB
• 2: Module: TigerBeetle
• 3: Module: Kubernetes
• 4: Module: Consul

1 - Module: DuckDB

DuckDB is a high-performance embedded analytical database.

DuckDB is embedded, so it does not require deployment or service management. Install the DuckDB package on the node and use it directly.

Installation

The current Pigsty node platform mapping includes the duckdb package, which can be installed directly from the Infra repository:

./node.yml -t node_install  -e '{"node_repo_modules":"infra","node_packages":["duckdb"]}'

Install with pig:

pig repo add infra -u      # add Infra repo
pig install duckdb         # install DuckDB package

Resources

Pigsty provides DuckDB-related PostgreSQL extensions:

• pg_duckdb, the official DuckDB PostgreSQL extension
• pg_mooncake, builds on pg_duckdb with columnar engine and sync
• pg_analytics, high-performance OLAP built on DuckDB; the project is archived and has been removed from the current Pigsty extension catalog
• duckdb_fdw, a DuckDB foreign data wrapper for reading and writing DuckDB files from PostgreSQL

2 - Module: TigerBeetle

TigerBeetle is a financial accounting transaction database offering extreme performance and reliability.

Overview

The current open-source tree has no TigerBeetle role or dedicated playbook. It only provides the tigerbeetle installation alias in the node platform package map; initialize and manage the service according to the official TigerBeetle documentation.

Installation

Use the following command to install the mapped package from the Pigsty Infra repository:

./node.yml -t node_install -e '{"node_repo_modules":"infra","node_packages":["tigerbeetle"]}'

After installation, please refer to the official documentation for configuration: https://github.com/tigerbeetle/tigerbeetle

3 - Module: Kubernetes

Kubernetes is a production-grade, open-source container orchestration platform. It helps you automate, deploy, scale, and manage containerized applications.

Pigsty has native support for ETCD clusters, which can be used as external etcd for Kubernetes.

The current open-source tree has no kube.yml playbook, Kubernetes role, or kube_* parameters. Pigsty provides software repositories and base-node provisioning here; cluster initialization, networking, control-plane management, and upgrades must still be handled by kubeadm, SealOS, or another Kubernetes tool.

SealOS

SealOS is a lightweight, high-performance, and easy-to-use Kubernetes distribution. It is designed to simplify the deployment and management of Kubernetes clusters.

The current Pigsty node platform mapping includes the sealos package. Install it from the Infra repository, then use SealOS to manage the cluster.

./node.yml -t node_install -e '{"node_repo_modules":"infra","node_packages":["sealos"]}'

Kubernetes

If you use classic kubeadm to deploy Kubernetes, install these packages first:

./node.yml -t node_install -e '{"node_repo_modules":"kube","node_packages":["kubeadm","kubelet","kubectl"]}'

Kubernetes supports multiple container runtimes. If you want to use Containerd as the container runtime, please make sure Containerd is installed on the node.

./node.yml -t node_install -e '{"node_repo_modules":"node,infra","node_packages":["containerd.io"]}'

To use Docker as the container runtime, install Docker and provide the cri-dockerd bridge component yourself. The default package mapping currently only includes the containerd.io runtime:

./node.yml -t node_install -e '{"node_repo_modules":"node,infra","node_packages":["docker-ce","docker-compose-plugin"]}'

Monitoring

Kubernetes cluster observability is typically handled by in-cluster stacks (such as kube-prometheus-stack).

On the Pigsty side, you can monitor the foundational dependencies Kubernetes relies on:

• ETCD Monitoring & Alerting: Control-plane metadata consistency and availability
• NODE Monitoring & Alerting: Host-level CPU, memory, kernel, and network health
• INFRA Monitoring & Alerting: Monitoring backend, alert pipeline, and observability platform health

4 - Module: Consul

Consul is a distributed DCS, KV, DNS, and service registry/discovery component.

Pigsty 1.x used Consul as its high-availability DCS. The current open-source tree has removed the Consul role, playbook, and consul_* parameters, so the legacy inventory shown on older versions cannot be used to deploy it. Current Pigsty DCS integration is provided by the ETCD module.