Issue Summary

Deleting a node using the DELETE /api/2/cluster/node/{nodeId} REST API does not fully purge all metadata. While the node record is removed from the clusternode table, stale records remain in the clusternodeheartbeat table.

Steps to Reproduce

1. Setup a 2-node Jira Data Center clustered environment.
2. Shut down one of the nodes.
3. Once the node is reported as "Offline", delete it using the REST API:

   curl -X DELETE -u $ADMIN_USER "$BASE_URL/rest/api/2/cluster/node/$NODE_ID"
   

4. Check the database tables clusternode and clusternodeheartbeat.

Expected Results

The records for $NODE_ID should be purged from both the clusternode and clusternodeheartbeat tables to ensure no stale metadata remains.

Actual Results

The record is removed from clusternode, but a stale entry persists in clusternodeheartbeat.

Workaround

Stop the entire cluster and manually delete the stale heartbeat record from the database:

DELETE FROM clusternodeheartbeat WHERE NODE_ID = '$NODE_ID';

Note: Always perform a database backup and test in a lower environment before executing direct SQL manipulations.