The MOVED error in Redis Cluster is a normal part of the redirection mechanism - it tells clients which node is responsible for handling a particular key. However, in your case, there appears to be a fundamental issue with your cluster configuration.

Looking at your cluster configuration:

192.168.0.14:6379 master (slots from 0 to 16383)
192.168.0.15:6379 slave (slots from 0 to 16383)
192.168.0.16:6379 master (without slots)

There are several critical issues:

1. 192.168.0.14 claims all 16384 slots (0-16383)
2. 192.168.0.16 has no assigned slots
3. This creates an unbalanced cluster where one node handles all traffic

When you execute:

192.168.0.16:6379> set myKey myValue
(error) MOVED 16281 192.168.0.14:6379

The cluster is working correctly by redirecting you to the proper node. However, the command appears to fail because:

1. The client isn't automatically following the redirect
2. The key might not actually be stored (shown by the nil response)

Here's how to properly set up a 3-node cluster with balanced slots:

# On each node
redis-cli --cluster create \
192.168.0.14:6379 \
192.168.0.15:6379 \
192.168.0.16:6379 \
--cluster-replicas 0

This will automatically distribute slots evenly across all masters.

For production use, you should:

1. Use a Redis client that supports cluster mode (like redis-py-cluster)
2. Or handle MOVED/ASK responses manually

Example Python code using redis-py-cluster:

from rediscluster import RedisCluster

startup_nodes = [{"host": "192.168.0.14", "port": "6379"}]
rc = RedisCluster(startup_nodes=startup_nodes, decode_responses=True)

# This will automatically handle redirections
rc.set("myKey", "myValue")
print(rc.get("myKey"))

Always check your cluster status with:

redis-cli --cluster check 192.168.0.14:6379

This will show slot distribution and any configuration issues.

• Redis Cluster requires proper slot distribution across masters
• Clients must either follow redirects or use cluster-aware libraries
• Never have a master with no slots assigned
• Regularly check cluster health with cluster check

---

When working with Redis Cluster, you'll encounter the MOVED error when attempting to access keys that don't belong to the current node's slot range. The error follows this format:

(error) MOVED [slot] [destination-node-ip:port]

In your specific case, the cluster configuration shows:

192.168.0.14:6379 master (slots from 0 to 16383)
192.168.0.15:6379 slave (slots from 0 to 16383)
192.168.0.16:6379 master (without slots)

The key issue here is that node 192.168.0.16 has no assigned slots, making it unable to process any key operations directly.

When you execute:

192.168.0.16:6379> set myKey myValue
(error) MOVED 16281 192.168.0.14:6379

The cluster correctly identifies that the key's slot (16281) belongs to 192.168.0.14, but the command doesn't automatically follow the redirection because:

• Redis CLI in standalone mode doesn't handle cluster redirections by default
• The target node (192.168.0.14) returns (nil) because the initial SET operation never reached it

Option 1: Using Redis CLI Cluster Mode

Connect using cluster-aware client mode:

redis-cli -c -h 192.168.0.16 -p 6379

Now operations will automatically follow redirections:

192.168.0.16:6379> set myKey myValue
-> Redirected to slot [16281] located at 192.168.0.14:6379
OK
192.168.0.14:6379> get myKey
"myValue"

Option 2: Proper Cluster Configuration

Your cluster has unbalanced slot distribution. Fix this by:

redis-cli --cluster reshard 192.168.0.14:6379
# Follow prompts to move slots to 192.168.0.16

After resharding, check cluster status:

redis-cli --cluster check 192.168.0.14:6379

Option 3: Programming Client Handling

For application code, use cluster-aware clients. Example in Python:

from rediscluster import RedisCluster

startup_nodes = [{"host": "192.168.0.14", "port": "6379"}]
rc = RedisCluster(startup_nodes=startup_nodes, decode_responses=True)
rc.set("myKey", "myValue")  # Automatically handles redirections

• Always use -c flag for cluster-aware CLI connections
• Ensure proper slot distribution across all master nodes
• Use cluster-enabled client libraries in production applications
• The (nil) response indicates the initial operation never reached the correct node