[pgpool-general: 8573] PgPool thinks node 0 is in recovery.

Ron ronljohnsonjr at gmail.com
Fri Jan 27 05:06:51 JST 2023


psql on FISPCCPGS405a port 5432 says that pg_is_in_recovery() = 'f' (that is 
correct), while psql on FISPCCPGS405a port 9999 says that 
pg_is_in_recovery() = 't'.

That is incorrect, and certainly why "psql --host=localhost --port=9999" 
drops me into a read only transaction.

What should I look for (logging is debug3") to resolve this?

$ hostname
FISPCCPGS405a


$ grep FISPCCPGS405[ab] pgpool.conf
backend_hostname0 = 'FISPCCPGS405a'
backend_hostname1 = 'FISPCCPGS405b'
hostname0 = 'FISPCCPGS405a'
hostname1 = 'FISPCCPGS405a'


$ psql --host=localhost --username=postgres --port=9999 -c "\x" -c "show 
pool_nodes;"
Expanded display is on.
-[ RECORD 1 ]----------+--------------------
node_id                | 0
hostname               | FISPCCPGS405a
port                   | 5432
status                 | down <<<<<<<<<<<<<<
pg_status              | up <<<<<<<<<<<<<<
lb_weight              | 0.666667
role                   | primary
pg_role                | primary
select_cnt             | 0
load_balance_node      | false
replication_delay      | 0
replication_state      |
replication_sync_state |
last_status_change     | 2023-01-26 13:31:04
-[ RECORD 2 ]----------+--------------------
node_id                | 1
hostname               | FISPCCPGS405b
port                   | 5432
status                 | up <<<<<<<<<<<<<<<
pg_status              | up <<<<<<<<<<<<<<<
lb_weight              | 0.333333
role                   | standby
pg_role                | standby
select_cnt             | 0
load_balance_node      | true
replication_delay      | 0
replication_state      |
replication_sync_state |
last_status_change     | 2023-01-26 13:31:04



-- 
Born in Arizona, moved to Babylonia.


More information about the pgpool-general mailing list