pcp_detach_node detaches the given node from Pgpool-II. If failover_command and/or follow_primary_command are specified, they are executed too. Existing connections to Pgpool-II are forced to be disconnected.
pcp_detach_node just detaches the node, and does not touch running backend behind the node. This command is useful when admin needs to maintain the PostgreSQL node. He/she can shutdown or stop the backend as many times as he/she wants.
The safest way to re-attach the detached node is, stopping the backend and apply pcp_recovery_node. However if you are sure that there's no replication delay (or the delay will be recovered later on) and the role of the node (primary/standby) will not be changed, you can use pcp_attach_node.
The index of backend node to detach.
wait until all clients are disconnected (unless client_idle_limit_in_recovery is -1 or recovery_timeout is expired).
See pcp_common_options.