Manage namespaces

Aerospike Database clusters contain one or more namespaces, similar to tablespaces in an RDBMS. Namespaces segregate data with different storage requirements, so you can optimize performance by keeping different kinds of data on different kinds of storage hardware.

Namespaces are defined in the main Aerospike configuration file. You can add, remove, or rename namespaces by editing aerospike.conf. You can configure namespaces for optimal performance in a variety of use cases by specifying a storage engine, a data retention policy, and other parameters.

The Aerospike server reads the configuration file when the server starts. You need to stop the server, edit the configuration file, and restart the server for changes to take effect. You can perform a rolling restart of cluster nodes to prevent server downtime.

We recommend that you back up your cluster before performing namespace alterations.

Minimum namespace configuration

The minimum requirement to configure a namespace is to provide it with a name. The following example creates an in-memory namespace with a 4GB capacity. The commented parameters indicate the defaults.

namespace <namespace-name> {
    # memory-size 4G           # 4GB of memory to be used for index and data
    # replication-factor 2     # For multiple nodes, keep 2 copies of the data
    # high-water-memory-pct 60 # Evict non-zero TTL data if capacity exceeds
                               # 60% of 4GB
    # stop-writes-pct 90       # Stop writes if capacity exceeds 90% of 4GB
    # default-ttl 0            # Writes from client that do not provide a TTL
                               # will default to 0 or never expire
    # storage-engine memory    # Store data in memory only
}

Add or remove a namespace

Perform the following steps on the cluster nodes one at a time:

1. Stop the Aerospike service

   systemctl stop aerospike

2. Update the aerospike.conf file

   Add or remove namespace contexts as needed.

3. Start the Aerospike service**

   systemctl start aerospike

Rename an existing namespace

It’s not possible to rename an existing namespace in place, but you can create a new namespace with the desired name and move data from an existing namespace into it.

1. Stop application traffic for the namespace to be renamed

2. Back up the namespace to be renamed

   You can use asbackup and asrestore for on-demand cluster backup.

   Execute the following steps on the cluster nodes one at a time:

   1. Stop the Aerospike service

      systemctl stop aerospike

   2. Delete stored data for the namespace to be renamed

      If your data is stored in a file, delete the file. For raw devices, use the blkdiscard or dd Linux utilities. The Zeroize Multiple SSDs Simultaneously Knowledge Base article has more information about deleting data from raw devices.

   3. Modify aerospike.conf to update the namespace context.

   4. Start the Aerospike service

      systemctl start aerospike

   5. Restore the old namespace’s data

      When all the cluster nodes have restarted with the new namespace configuration, use the -n flag to to restore the data and specify the new namespace.

When the restore process is complete, the new namespace is ready to use.

Related topics

• Storage engine configures if records are persisted and where.
• Configure a data retention policy to determine how long records are kept after initially written.
• Data durability determines how many replica copies of a record to keep in the cluster.
• Configure strong consistency describes how to set up SC mode for transactions.
• For a complete list of namespace configuration parameters see the Configuration reference.