Performing the Redis Software 8.0.2 Upgrade (In-Place and Rolling Methods)

Redis Software 8.0.2 introduces validated upgrade paths, new operational safeguards, and compatibility changes across clustered, HA, and Kubernetes deployments. This guide provides supported upgrade paths, required pre-checks, step-by-step instructions for both in-place and rolling upgrades, post-upgrade validation steps, and troubleshooting for common failures.

Sections covered include: Prerequisites, Upgrade Methods, Post-Upgrade Actions, Troubleshooting, and Kubernetes Considerations.

Prerequisites

Supported Upgrade Paths

Kubernetes: Only upgrades from Operator 7.4.2-2 or later are supported. See the Kubernetes 8.0.2-2 Release Notes.

Required System and Cluster Checks

Review release notes

• Read the RS 8.0.2-17 Release Notes to understand new features, behavior changes, and known limitations.

Confirm your upgrade path

• Verify your current RS version supports a direct upgrade using the Supported Upgrade Paths table.

Check database compatibility

• Databases must run Redis 6.2, 7.2, 7.4, 8.0, or 8.2.
• Upgrade any database older than 6.2 before upgrading the cluster.

Verify node access

• Ensure rlcheck and rladmin can be executed as root on every node.
• Confirm SSH or console access to all machines.

Validate cluster health

• Run:
  • rladmin status extra all
  • rladmin cluster running_actions
• All shards must be healthy and no running actions should be present.

Identify the primary node

• Run rladmin status nodes to locate the primary/master node.
• Always upgrade the primary node first.

Perform and validate backups

• Create full backups for every database.
• Perform a test restore on a staging or validation cluster.

Freeze topology changes

• Pause migrations, scaling, configuration changes, module updates, or database edits until the entire upgrade is complete.

Upgrade Methods

Choose the method based on cluster size, environment, and availability requirements.

In-Place Upgrade (Per Node)

Best for test and non-HA environments.

1. Validate node health

• Run rlcheck
• Run rladmin status extra all
  Proceed only when all checks pass.

2. Download RS 8.0.2 installation package

• Obtain from the Redis Download Center.

3. Stop external integrations

• Disable monitoring or custom automation interacting with Redis.

4. Install as root:

sudo ./install.sh

5. Verify post-install state

• rlcheck
• rladmin status

6. Repeat for each node
Upgrade one node at a time and wait for the system to report OK before moving on.

Notes

• Do not modify installation directories or service users.
• Ensure /opt/redislabs/bin remains in PATH.

Rolling Upgrade Methods

Recommended for production deployments requiring high availability.

Extra Node Method (Preferred)

1. Add a new RS 8.0.2 node
Install RS 8.0.2 on a clean machine.

2. Join it to the cluster:

rladmin cluster join

3. (Optional) Promote to primary
If replacing the master.

4. Remove an old-version node:

rladmin cluster remove node <node_id>

5. Repeat
Replace nodes one at a time until all are running RS 8.0.2.

Replace Node Method

Used when no extra hardware is available.

1. Remove old node from cluster:

rladmin cluster remove node <node_id>

2. Uninstall Redis if necessary:

sudo ./rl_uninstall.sh

3. Install RS 8.0.2:

sudo ./install.sh

4. Rejoin via replace-node:

rladmin cluster join nodes <cluster_ip> \
username <username> password <password> replace_node <node_id>

5. Validate:

• rlcheck
• rladmin status extra all

Repeat until all nodes are upgraded.

Post-Upgrade Actions

For Active-Active (CRDB)

Upgrade cluster instances before upgrading databases.

1. Sync module configurations:

crdb-cli crdb update --crdb-guid <guid> --update-db-config-modules true

2. Upgrade each database:

rladmin upgrade db db:<id> preserve_roles latest_with_modules

3. Accept protocol upgrade when prompted.

Troubleshooting and Common Issues

Quick Fix Table

Detailed Troubleshooting

Background tasks stuck

• Check for and clear running actions:
  • rladmin cluster running_actions

Port conflicts

• Stop any lingering Redis processes (ps -ef | grep redis)
• Restart RL services.

Shard migration failures

• Migrations are supported only after all nodes run RS 8.0.2.

PATH lost after install

• Re-add /opt/redislabs/bin to PATH if update_env_path=no was used.

Active-Active module mismatch

• Wait until all participating clusters and modules are upgraded.

Kubernetes Considerations

• Only upgrades from Operator 7.4.2-2 or later are supported.
• Custom modules are not supported in 8.0.2-2.
• Internode encryption uses default certificates only.
• Follow the steps in the Kubernetes 8.0.2-2 Release Notes for cluster upgrades.

References 

• RS 8.0.2 Release Notes
• Upgrade Redis Software Cluster 
• Supported Upgrade Paths for Redis Software
• Upgrade a Redis Enterprise Database
• Kubernetes 8.0.2-2 Release Notes