failback.sh

echo failover.sh: start: failed_node_id=$FAILED_NODE_ID failed_host=$FAILED_NODE_HOST 
    old_primary_node_id=$OLD_PRIMARY_NODE_ID new_main_node_id=$NEW_MAIN_NODE_ID new_main_host=$NEW_MAIN_NODE_HOST

## If there's no main node anymore, skip failover.
if [ $NEW_MAIN_NODE_ID -lt 0 ]; then
    echo failover.sh: All nodes are down. Skipping failover.
    exit 0
fi

## Test passwordless SSH
ssh -T ${SSH_OPTIONS} ${POSTGRESQL_STARTUP_USER}@${NEW_MAIN_NODE_HOST} ls /tmp > /dev/null

if [ $? -ne 0 ]; then
    echo failover.sh: passwordless SSH to ${POSTGRESQL_STARTUP_USER}@${NEW_MAIN_NODE_HOST} failed. Please setup passwordless SSH.
    exit 1
fi

## If Standby node is down, skip failover.
if [ $OLD_PRIMARY_NODE_ID != "-1" -a $FAILED_NODE_ID != $OLD_PRIMARY_NODE_ID ]; then

    # If Standby node is down, drop replication slot.
    ${PGHOME}/bin/psql -h ${OLD_PRIMARY_NODE_HOST} -p ${OLD_PRIMARY_NODE_PORT} postgres 
        -c "SELECT pg_drop_replication_slot('${REPL_SLOT_NAME}');"  >/dev/null 2>&1

    if [ $? -ne 0 ]; then
        echo ERROR: failover.sh: drop replication slot "${REPL_SLOT_NAME}" failed. You may need to drop replication slot manually.
    fi

    echo failover.sh: end: standby node is down. Skipping failover.
    exit 0
fi

## Promote Standby node.
echo failover.sh: primary node is down, promote new_main_node_id=$NEW_MAIN_NODE_ID on ${NEW_MAIN_NODE_HOST}.

ssh -T ${SSH_OPTIONS} ${POSTGRESQL_STARTUP_USER}@${NEW_MAIN_NODE_HOST} ${PGHOME}/bin/pg_ctl -D ${NEW_MAIN_NODE_PGDATA} -w promote

echo "primary_conninfo = 'user=repuser passfile=''/home/postgres/.pgpass'' channel_binding=prefer host=${NEW_MAIN_NODE_HOST} port=5432 sslmode=prefer sslcompression=0 sslsni=1 ssl_min_protocol_version=TLSv1.2 gssencmode=prefer krbsrvname=postgres target_session_attrs=any'" >> ${FAILED_NODE_PGDATA}/postgresql.auto.conf

# Create standby.signal file to indicate it's a standby node
touch ${FAILED_NODE_PGDATA}/standby.signal

if [ $? -ne 0 ]; then
    echo ERROR: failover.sh: end: failover failed
    exit 1
fi

echo failover.sh: end: new_main_node_id=$NEW_MAIN_NODE_ID on ${NEW_MAIN_NODE_HOST} is promoted to a primary
exit 0

I modified it based on examples. Except for “primary_conninfo” and the part that builds “standby.signal”

when i stop and start postgresql ，the old primary status became standby , stream replication still work

but when i check pgpool => show pool_nodes;

enter image description here

pg-status is up ，but status down
log does not show any error message，What happened? How do I fix it?

did i miss any config?Also, I would like to ask when should failback be used?