Redis Backups¶

This guide explains how to create Redis database backups, use them for local development, and restore them when needed. These procedures are essential for testing changes safely and performing rollbacks in production environments.

Prerequisites¶

redis-tools package (for redis-cli)
kubectl (for Kubernetes operations)
Docker (for local Redis instance)

Creating Redis Backups¶

There are two primary methods to create Redis backups:

Direct method: Connect to the Redis database using redis-cli and generate a backup directly on your local machine
Kubernetes method: Create a backup within the Redis pod and copy it to your local machine

Method 1: Using `redis-cli` Directly¶

This method allows you to create a backup directly from any Redis instance without requiring access to the underlying pod.

Step 1: Install Redis CLI Tools¶

# On Ubuntu/Debian
sudo apt-get install redis-tools

Step 2: Create the Redis Dump File¶

The following command connects to your Redis instance and creates a .rdb dump file containing all your data:

Creating a Redis dump file

redis-cli -a <redis_pwd> --rdb dump.rdb

Replace <redis_pwd> with your actual Redis password. This command will create a dump.rdb file in your current directory containing a complete snapshot of your Redis database.

Method 2: Using Kubernetes to Access a Redis Instance¶

This method is useful when you need to backup Redis instances running in Kubernetes clusters.

Step 1: Set Up Environment Variables¶

First, define a descriptive name for your backup file:

Setting up backup environment variables

export CSM_BACKUP_DUMP_FILE="<some-platform>-<some-tenant-name>-backup-example.rdb"

Replace <some-platform> and <some-tenant-name> with appropriate values that identify your environment.

Step 2: Create a Backup in the Redis Pod¶

These commands create a backup directory in the Redis pod and copy the current Redis dump file to it:

Creating a backup in the Redis pod

kubectl -n <some-namespace> exec --stdin --tty <some-pod> -- mkdir -p /opt/backups
kubectl -n <some-namespace> exec --stdin --tty <some-pod> -- cp /data/dump.rdb /opt/backups/$CSM_BACKUP_DUMP_FILE

Replace:

<some-namespace> with your Kubernetes namespace
<some-pod> with the name of your Redis pod

Step 3: Download the Backup File to Your Local Machine¶

After creating the backup in the pod, download it to your local machine:

Downloading the backup file

kubectl cp <some-namespace>/<some-pod>:/opt/backups/$CSM_BACKUP_DUMP_FILE /data/$CSM_BACKUP_DUMP_FILE
cp /data/$CSM_BACKUP_DUMP_FILE /data/dump.rdb

This copies the backup file from the pod to your /data directory and creates a copy named dump.rdb for easier reference.

Using a Redis Backup for Local Development¶

You can run a local Redis instance with your backup data for testing and development purposes.

Step 1: Pull the Redis Docker Image¶

First, pull the Cosmotech Redis image and tag it for easier reference:

Pulling the Redis Docker image

docker pull ghcr.io/cosmo-tech/cosmotech-redis:1.0.12
docker tag ghcr.io/cosmo-tech/cosmotech-redis:1.0.12 csm-redis

Step 2: Start a Local Redis Instance with Your Backup Data¶

Launch a Redis container using your backup data:

Starting a local Redis with backup data

docker run -e REDIS_PASSWORD=$REDIS_PASSWORD -e REDIS_AOF_ENABLED=no -p 6379:6379 -v /data:/bitnami/redis/data csm-redis /opt/bitnami/scripts/redis/run.sh --cluster-enabled no

This command:

Uses the environment variable REDIS_PASSWORD for authentication
Disables AOF persistence with REDIS_AOF_ENABLED=no
Maps port 6379 for Redis access
Mounts your local /data directory (containing the dump file) to the container's data directory
Runs the redis instance with cluster mode set to off, allowing a standalone local use

Once running, you can connect to this Redis instance using redis-cli with the password defined in the REDIS_PASSWORD environment variable.

Restoring a Backup to a Remote Redis Instance¶

This procedure allows you to restore a backup to a production or staging Redis instance.

Step 1: Disable Redis Persistence on the Target Instance¶

First, connect to the Redis master instance and disable the save operation to prevent data corruption during restore:

Disabling save on the Redis instance

kubectl -n <some-namespace> exec --stdin --tty <name-master-redis-instance> -- /bin/bash
redis-cli -a $REDIS_PASSWORD
CONFIG set save ""
exit
exit

This sequence:

Opens a shell in the Redis master pod
Connects to Redis using the CLI
Disables the automatic save operation
Exits both the Redis CLI and the pod shell

Step 2: Upload and Apply the Backup File¶

Copy your local backup file to the Redis master pod and replace the current dump file:

Uploading the backup file to the pod

kubectl cp $PWD/data/$CSM_BACKUP_DUMP_FILE <some-namespace>/<name-master-redis-instance>:/data/dump.rdb-1
kubectl -n <some-namespace> exec --stdin --tty <name-master-redis-instance> -- chmod 777 /data/dump.rdb-1
kubectl -n <some-namespace> exec --stdin --tty <name-master-redis-instance> -- cp /data/dump.rdb-1 /data/dump.rdb
kubectl -n <some-namespace> exec --stdin --tty <name-master-redis-instance> -- rm /data/dump.rdb-1