Skip to main content

Data Rebalancing

danger

This is an advanced operation. Please make sure to have a Forward Support Engineer on a call when performing this operation to avoid mistakes that can cause data loss.

In multi-node deployments, it is possible that the disk on some nodes is mostly used but some other nodes have a lot of free disk space. It would be desirable to move some database replicas out of the heavily used nodes to better rebalance the disk usage across cluster.

Forward provides a helper script for this purpose. This script verifies several pre-conditions before allowing a database replica to be migrated to ensure we don't encounter data loss.

Steps to migrate a database replica:​

  • SSH into a primary node

  • Launch the fwd-cluster utility and execute the below command:

    Forward Cluster> db migrate

  • From the list enter the database replica pod name to migrate

Location Map

  • Select the node which doesn’t have not eligible as suffix. This node would be the target to move the database replica.

Location Map

  • Then the script will perform the migration.

Location Map

  • Once the migration is finished, execute the below command to verify the database replica distribution:

    Forward Cluster> db status

Location Map