Admin guide overview
This guide shows you how to create, run, and manage Clustron Zaris stores from the AdminShell PowerShell module. It covers the day-to-day operational tasks: connecting to managers, creating and starting stores, running across multiple nodes, understanding ports, and monitoring and troubleshooting a live cluster.
A store is a distributed key-value database made up of one or more nodes. You administer a store through managers — control-plane services that create the nodes and start, stop, and report on them. You never edit node configuration files directly; you issue cmdlets, and the managers apply the change across the cluster.
The administrative workflow
Managing a store follows the same order every time. Each step depends on the one before it.
- Activate a workspace with
New-ZrWorkspace <name> @('host:7801') -Activate(orUse-ZrWorkspacefor one that already exists). Without-Activatethe workspace is only saved, not made active. This points every other admin cmdlet at the cluster's managers. - Create the store with
New-ZrStore. This registers the node topology but does not start anything. - Start the store with
Start-ZrStore. This launches the nodes and forms the cluster. - Verify and monitor with
Get-ZrStoreandWatch-ZrStoreMetrics. - Stop the store with
Stop-ZrStorewhen you no longer need it.
Managers and nodes
Clustron separates the control plane from the data plane.
- A manager runs on each server, listens on the management port (default
7801), and controls the stores on that server. It creates, starts, and stops nodes, and reports their status. - A node (also called a store instance) is a running process that holds data and coordinates with the other nodes in the store. Each node listens on its own cluster port and client port.
Server
├── Manager (7801)
└── Node(s) (cluster port + client port each)
One manager runs per server. In a cluster that spans several servers, you connect to all of the managers at once so that every administrative change is applied consistently across the whole cluster.
The AdminShell cmdlets
You administer Zaris with the Clustron.Zaris.AdminShell PowerShell module. Import it once per session:
Import-Module Clustron.Zaris.AdminShell
The core cmdlets map directly to the workflow above.
| Cmdlet | Purpose |
|---|---|
New-ZrWorkspace <name> @('host:7801') -Activate | Create and activate a workspace targeting one or more managers (-Activate makes it the session's target; omit it to only save) |
Use-ZrWorkspace | Reactivate or join an existing workspace |
New-ZrStore | Create a store and its node topology |
Start-ZrStore | Start all nodes of a store, or one named node |
Stop-ZrStore | Stop all nodes of a store, or one named node |
Get-ZrStore | Report store status, node counts, and health |
Watch-ZrStoreMetrics | Show live per-node throughput and latency |
Single-server and multi-server topologies
You can run Zaris on one server or across many. The topology you choose determines how you set -PartitionsPerNode and -ReplicationFactor when you create the store.
- Single server, multiple nodes. One manager runs several node processes on the same machine, controlled with
-PartitionsPerNode(and the replication factor). Use this for development and testing. - Multiple servers. One manager runs on each server, and each server runs its share of the partitions and replicas. Use this for production, higher capacity, and fault tolerance.
Both models are covered in detail in Multi-node configuration.
Ports at a glance
A Zaris server divides the 7801–7899 span into three ranges, one per kind of traffic. The installer opens the whole span in the local firewall.
- Management (
7801–7810) — administrators, the manager-to-manager control plane, and the Web Console (7810). - Cluster (
7811–7850) — node-to-node coordination and data movement, from-BaseClusterPort(default7811). - Client (
7861–7899) — where applications and the ClientShell connect, from-BaseClientPort(default7861).
See Ports and networking for the full ranges and firewall guidance.
Next steps
- Connect to a cluster — establish the manager session.
- Create a store — define the node topology.
- Ports and networking — the port ranges and firewall rules.
- Performance and benchmarking — measure and tune throughput and latency.
- Capacity planning — size the cluster for your working set.
- High availability and disaster recovery — survive node loss.
- Upgrades and lifecycle — restart, scale, and upgrade safely.