Redis Software uses a shared-nothing architecture, where each database runs in isolation with its own memory, storage, and network resources. All client connections are handled through the Redis proxy service, and each database is assigned a unique host and port combination at creation. This guide explains how to identify the correct host and port using the Admin Console, CLI tools, and cluster design constraints. It also outlines common access issues and clarifies architecture-specific behaviors such as proxy routing and port reservation. See Step-by-Step Instructions, Troubleshooting, and Behavioral Clarifications for full details.
Key Architecture Notes
Proxy-Based Access Only: All client connections go through the Redis proxy. Even with the OSS Cluster API enabled, traffic still flows via proxy.
No redis.conf Edits: Redis Software manages all configuration at the cluster and database level. Manual configuration changes (e.g., redis.conf) are not supported and not applicable.
-
Reserved Ports Only: Ports must be selected from the Redis Software reserved port ranges:
10000–1004910051–19999
Port assignment happens during database creation and cannot be changed afterward.
Admin Console Port (8443): Used only to access the web UI. It is not a database access port.
Step-by-Step Instructions
Identify Host and Port via Admin Console
Open the Admin Console:
https://<cluster_fqdn>:8443Navigate to Databases → [Your Database]
-
Under the General or Endpoint section, note:
Database Endpoint (Host + Port) — e.g.,
redis-10269.cluster-name.example.com:10269Public Endpoint—visible only if public access was enabled at cluster creation
Identify Endpoint via CLI (rladmin)
From a Redis cluster node, run:
rladmin status databases
Example output:
DB:ID NAME TYPE STATUS SHARDS PLACEMENT REPLICATION PERSISTENCE ENDPOINT db:1 database1 redis active 1 dense disabled disabled redis-10269.testdbd11169.localhost:10269 db:2 database2 redis active 1 dense disabled snapshot redis-13897.testdbd11169.localhost:13897 db:3 database3 redis active 1 dense enabled snapshot redis-19416.testdbd13186.localhost:19416
The ENDPOINT column provides the correct host and port for client use.
All traffic to this endpoint is automatically routed through the Redis proxy.
Common Issues and Troubleshooting
Issue |
Cause |
Solution |
|---|---|---|
Host unreachable |
Network segmentation or DNS failure |
Validate DNS, VPC/subnet configuration, and firewall rules |
Port inaccessible |
Inbound rules not allowing selected DB port |
Ensure the chosen port is open within your security groups/firewalls |
TLS-related failures |
Incorrect client scheme (e.g. |
Use |
Database unavailable |
Database not yet provisioned or restarted |
Verify DB status in Admin Console or via |
Behavioral Clarifications
All connections route through proxy
Clients must use the proxy-managed hostname and port.
Direct shard/node IP connections are not permitted in Redis Software.
Default and Custom Port Usage
Redis Software does not use 6379 or 6380 (those are Community defaults).
Ports must be selected from reserved ranges. Custom arbitrary ports are not supported.
Standalone vs OSS Cluster API
Clients should use standalone connection mode unless the OSS Cluster API is explicitly enabled.
Cluster-aware clients (e.g., using
CLUSTER SLOTS) will fail unless OSS Cluster mode is turned on.
Public Endpoints
Public endpoints are only available if enabled during cluster creation.
By default, databases are accessible only over internal networking.
Active-Active (CRDB) Considerations
Each participating region in a CRDB deployment will expose its own proxy endpoint.
Clients should connect to the local region’s proxy for lowest-latency access.
Additional Resources
0 comments
Please sign in to leave a comment.