Query blocked when one node is down

I follow the instruction to start a citus cluster with 6 node.
I created distributed table with 2 replications.

I was expecting the cluster to work after I shutdown one worker node.
But actually, if I use docker stop to stop a worker, the simple select * from test_table will block.

I thought citus as high availability natively, am I wrong?