Environment

• YugabyteDB Anywhere < 2.16.2

Issue

In earlier versions of YugabyteDB Anywhere, when a node was released to the node pool, data on the node and yugabyte directories were not properly deleted. When a previously released node was re-added to the cluster, the cluster does not properly bootstrap the new node since the data and yugabyte directories already exist.

Resolution

Overview

To ensure than an on-prem node is safe to add to the cluster, manually perform the cleanup steps to ensure the relevant directories are deleted.

Note: This guide assumes you have already removed the node from the cluster and released the instance back to the node pool

Steps

1. SSH in to the removed node as a user with sudo or root permissions.

2. Manually wipe the disks.

Note: This command assumes that your data directory is in the default location of mnt/dX

sudo rm -rf /mnt/d*/

3. On YugabyteDB Anywhere In the Actions > Edit Universe dropdown, use these documentation steps to add a node to the universe  

Additional Information

We are deleting the data on the disks so that if the node is added to the cluster again, it will be prepared as a brand new node in the cluster