Upgrade a HarperDB Instance

This document describes best practices for upgrading self-hosted HarperDB instances. HarperDB can be upgraded using a combination of npm and built-in HarperDB upgrade scripts. Whenever upgrading your HarperDB installation it is recommended you make a backup of your data first. Note: This document applies to self-hosted HarperDB instances only. All HarperDB Cloud instances will be upgraded by the HarperDB Cloud team.

Upgrading

Upgrading HarperDB is a two-step process. First the latest version of HarperDB must be downloaded from npm, then the HarperDB upgrade scripts will be utilized to ensure the newest features are available on the system.

1. Install the latest version of HarperDB using npm install -g harperdb.

   Note -g should only be used if you installed HarperDB globally (which is recommended).

2. Run harperdb to initiate the upgrade process.

   HarperDB will then prompt you for all appropriate inputs and then run the upgrade directives.

Node Version Manager (nvm)

Node Version Manager (nvm) is an easy way to install, remove, and switch between different versions of Node.js as required by various applications. More information, including directions on installing nvm can be found here: https://nvm.sh/.

HarperDB supports Node.js versions 14.0.0 and higher, however, please check our NPM page for our recommended Node.js version. To install a different version of Node.js with nvm, run the command:

nvm install <the node version>

To switch to a version of Node run:

nvm use <the node version>

To see the current running version of Node run:

node --version

With a handful of different versions of Node.js installed, run nvm with the ls argument to list out all installed versions:

nvm ls

When upgrading HarperDB, we recommend also upgrading your Node version. Here we assume you're running on an older version of Node; the execution may look like this:

Switch to the older version of Node that HarperDB is running on (if it is not the current version):

nvm use 14.19.0

Make sure HarperDB is not running:

harperdb stop

Uninstall HarperDB. Note, this step is not required, but will clean up old artifacts of HarperDB. We recommend removing all other HarperDB installations to ensure the most recent version is always running.

npm uninstall -g harperdb

Switch to the newer version of Node:

nvm use <the node version>

Install HarperDB globally

npm install -g harperdb

Run the upgrade script

harperdb

Start HarperDB

harperdb start

---

Upgrading Nats to Plexus 4.4

To upgrade from NATS clustering to Plexus replication, follow these manual steps. They are designed for a fully replicating cluster to ensure minimal disruption during the upgrade process.

The core of this upgrade is the bridge node. This node will run both NATS and Plexus simultaneously, ensuring that transactions are relayed between the two systems during the transition. The bridge node is crucial in preventing any replication downtime, as it will handle transactions from NATS nodes to Plexus nodes and vice versa.

Enabling Plexus

To enable Plexus on a node that is already running NATS, you will need to update two values in the harperdb-config.yaml file:

replication:
    url: wss://my-cluster-node-1:9925
    hostname: node-1

replication.url – This should be set to the URL of the current HarperDB instance.

replication.hostname – Since we are upgrading from NATS, this value should match the clustering.nodeName of the current instance.

Upgrade Steps

1. Set up the bridge node:

   • Choose one node to be the bridge node.

   • On this node, follow the "Enabling Plexus" steps from the previous section, but do not disable NATS clustering on this instance.

   • Stop the instance and perform the upgrade.

   • Start the instance. This node should now be running both Plexus and NATS.

2. Upgrade a node:

   • Choose a node that needs upgrading and enable Plexus by following the "Enable Plexus" steps.

   • Disable NATS by setting clustering.enabled to false.

   • Stop the instance and upgrade it.

   • Start the instance.

   • Call add_node on the upgraded instance. In this call, omit subscriptions so that a fully replicating cluster is built. The target node for this call should be the bridge node. Note: depending on your setup, you may need to expand this add_node call to include authorization and/or tls information.

{
    "operation": "add_node",
    "hostname:": "node-1",
    "url": "wss://my-cluster-node-1:9925"
}

1. Repeat Step 2 on all remaining nodes that need to be upgraded.

2. Disable NATS on the bridge node by setting clustering.enabled to false and restart the instance.

Your cluster upgrade should now be complete, with no NATS processes running on any of the nodes.