RadiatorDB node removal failed

Logged when removal of a configured RadiatorDB node fails, including failures that occur after the requesting client disconnects. The node field identifies the affected node; the attached error gives the underlying cause.

Before dropping a node's storage, removal checks that every remaining node has applied the removed node's writes, because no other node can supply them afterwards. If that check does not finish within the removal timeout, removal stops and keeps the removed node's storage; the error names the lagging node, its diagnosed replication worker state, and the next step. Removal also stops when dropping a subscription on a reachable node waits for a lock until the removal timeout expires; the error names the node, the subscription, and the step that waited.

Check the attached error and the node's lifecycle status in Radiator UI. Resolve the reported database, lock, replication, or storage problem before retrying Remove Node. If catch-up or lock waits time out, look for long-running transactions or lock waits on the named PostgreSQL node before adjusting RADIATORDB_NODE_REMOVAL_TIMEOUT. If a node is unreachable or a remaining node cannot catch up, forced removal removes the node at the cost of losing its unreplicated writes.