When a node is being shutdown for replacement, shards are only allowed to be copied from the source node to the destination replacement node. Once the migration of those shards is complete, the new node should be able to begin receiving additional shards. However, the allocation decider currently only looks at a replacement shutdown existing. It does not calculate the overall status of the shutdown, so even when the shutdown is complete, as long as the shutdown record exists additional allocations cannot be made to the new node.

While node shutdown records should be cleaned up, we shouldn't block allocation once the shutdown is complete.