ALTER CLUSTER DROP

Overview

Modifies an Xpand deployment, triggering a group change to explicitly removing a failed node from the deployment.

When using the Xpand Storage Engine topology, the details described here only apply when you connect to the Xpand nodes.

DETAILS

ALTER CLUSTER DROP nodeid

This statement starts a group change to immediately remove an Xpand Node from the deployment. Xpand does not wait for the rebalancer process to move data off the node, which can leave your deployment vulnerable to data loss.

Note

When a node is lost due to failure or ALTER CLUSTER DROP, the rebalancer processor attempts in the background to replace the lost replicas from existing copies to reprotect the deployment. If another node fails during this process, it can result in partial availability.

Only use ALTER CLUSTER DROP in the event of emergency.

To remove a node without risking data loss, use ALTER CLUSTER SOFTFAIL instead.

Removing Nodes

To remove a node:

  1. Connect to an existing Xpand Node with the socket file used by the MariaDB Xpand service:

    $ mariadb --socket /data/clustrix/mysql.sock
    
  2. Issue an ALTER CLUSTER DROP statement to remove the node:

    ALTER CLUSTER DROP 4;
    

The Xpand Node and all replicas stored on it are no longer available to the deployment.