EMF Cloud Upgrade Guide#

Upgrade Path: EMF Cloud v3.1 → v25.2 Document Version: 1.0

Overview#

This document provides step-by-step instructions to upgrade Cloud Edge Manageability Framework (EMF) from version 3.1 to 25.2.

Important Notes#

Warning

DISRUPTIVE UPGRADE WARNING This upgrade requires edge node re-onboarding due to architecture changes (RKE2 → K3s). Plan for edge nodes service downtime and manual data backup/restore procedures in edge nodes.

Prerequisites#

System Requirements#

Current EMF Cloud installation version 3.0
Root/sudo privileges on orchestrator node
PostgreSQL service running and accessible
docker user credential if any pull limit hit

Pre-Upgrade Checklist#

[ ] Back up critical application data from edge nodes
[ ] Document current edge node configurations
[ ] Remove all edge clusters and hosts: - Delete clusters - De-authorize hosts - Delete hosts

Upgrade Procedure#

Step 1: Prepare Edge Orchestrator Upgrade Environment#

You need to follow the steps mentioned in Prerequisites section with some changes as listed below.

Pull the intended 25.2 cloud installer image and extract it

# Replace <25.2-TAG> with actual tag
oras pull registry-rs.edgeorchestration.intel.com/edge-orch/common/files/cloud-orchestrator-installer:<25.2-TAG>
tar -xzf _build/cloud-orchestrator-installer.tgz

Start the installation Environment

Start the Edge Orchestrator installer

./start-orchestrator-install.sh

Type 2 for managing an existing cluster

Type in cluster details, including cluster name and the AWS region. (Same values as were passed during cluster provisioning)

Specify a location to store the installer settings. (Same values as were passed during cluster provisioning)

Provision the Environment

Go to pod-configs directory

orchestrator-admin:~$ cd ~/pod-configs

Configure the cluster provisioning parameters

orchestrator-admin:~/pod-configs$ ./utils/provision.sh config \
--aws-account [AWS account] \
--customer-state-prefix [S3 bucket name prefix to store provision state] \
--environment [Cluster name] \
--parent-domain [Root domain for deployment] \
--region [AWS region to install the cluster] \
--jumphost-ip-allow-list [IPs to permit cluster administration access]

Note

Follow the official guide to get details about each parameter

Run the following command to begin upgrade

orchestrator-admin:~/pod-configs$ ./utils/provision.sh upgrade \
--aws-account [AWS account] \
--customer-state-prefix [S3 bucket name prefix to store provision state] \
--environment [Cluster name] \
--parent-domain [Root domain for deployment] \
--region [AWS region to install the cluster] \
--jumphost-ip-allow-list [IPs to permit cluster administration access]

Upgrade Edge Orchestrator

Go to home directory
```
orchestrator-admin:~$ cd ~
```
Load the aws environment variables into the shell
source ./reconnect-aws-cluster.sh
Configure the cluster deployment options. From the ~ directory in the orchestrator-admin container, run the following command:
```
orchestrator-admin:~$ ./configure-cluster.sh
```
This process will start redeploying the upgraded applications in the cluster starting with root-app. Let it continue and you would observe “infra-external” app is failing due to orch-infra-rps and orch-infra-mps databases. In order to fix the above problem, you need to follow below steps.

Step 2: Verification#

Log into web UI of the orchestrator. Go to Settings->OS profiles. There you should see the any of the toolkit version upgraded to latest.

Post-Upgrade Steps Edgenode onboarding process#

After a successful upgrade, follow the EN onboarding process as outlined in the official documentation: Set Up Edge Infrastructure – Intel Open Edge Platform