What Are the Constraints and Precautions for Migrating Redis Data to a Cluster Instance?

Proxy Cluster instances are used in the same way that you use single-node or master/standby instances. However, only one DB is configured for a Proxy Cluster instance by default, and the SELECT command is not supported. When data files are imported in batches, an error message will be displayed and ignored if the SELECT command exists. Then, the remaining data will continue to be imported.

Example:

DB 0 and DB 2 in the source Redis instance contain data, and the generated AOF or RDB file contains these two DBs.

When the source Redis data is imported into a Proxy Cluster DCS instance, the SELECT 2 command will be ignored, and then data in DB 2 in the source Redis instance will be imported.

Note that:

• If different DBs in the source Redis instance contain the same keys, values of keys in the DB with the largest ID will overwrite those in the other DBs.
• If the source Redis instance contains multiple DBs, data is stored in the same DB after being migrated to a cluster DCS instance, and the SELECT command is not supported. In this case, modify the configurations for your services.

Only one DB is configured for a Redis Cluster instance. Data is migrated to a Redis Cluster instance in a different way from other types of instances. Nodes in the shards of a Redis Cluster must be connected separately through clients. Data is imported to the nodes separately. Run the following command to query the IP addresses of the cluster nodes:

redis-cli -h {Redis Cluster IP} -p 6379 -a {password} cluster nodes

In the returned list of IP addresses, record the ones marked by "master".