Hellmuth Vargas escribió:

> > Ahh, ¿no tendrás un standby con "feedback" activado que tiene una
> > transacción preparada, o un slot inactivo?  Pega acá
> > select * from pg_stat_replication
> 
> 
> bd=# select * from pg_stat_replication
> bd-# ;
> -[ RECORD 1 ]----+------------------------------
> pid              | 1892
> usesysid         | 10
> usename          | postgres
> application_name | walreceiver
> client_addr      | 192.168.72.XX
> client_hostname  |
> client_port      | 46648
> backend_start    | 2016-10-02 11:46:57.003511-05
> backend_xmin     |
> state            | streaming
> sent_location    | 9/11E181C0
> write_location   | 9/11E181C0
> flush_location   | 9/11E181C0
> replay_location  | 9/11E18090
> sync_priority    | 0
> sync_state       | async
> -[ RECORD 2 ]----+------------------------------
> pid              | 2043
> usesysid         | 10
> usename          | postgres
> application_name | walreceiver
> client_addr      | 10.72.73.YY
> client_hostname  |
> client_port      | 55716
> backend_start    | 2016-10-02 11:48:07.38359-05
> backend_xmin     |
> state            | streaming
> sent_location    | 9/11E181C0
> write_location   | 9/11E181C0
> flush_location   | 9/11E181C0
> replay_location  | 9/11E18090
> sync_priority    | 0
> sync_state       | async

???

Según controldata que pasaste antes,
  Latest checkpoint's REDO location:    1896/AB72AB00
No tiene ningún sentido que sent_location acá sea 9/11E181C0. ¿Estamos
hablando del mismo servidor?

> > select * from pg_replication_slots
> >
> > bd=# select * from pg_replication_slots;
> -[ RECORD 1 ]+---------------------
> slot_name    | replica_calleXX_slot
> plugin       |
> slot_type    | physical
> datoid       |
> database     |
> active       | t
> active_pid   | 1892
> xmin         | 8355084
> catalog_xmin |
> restart_lsn  | 9/11FA3D18
> -[ RECORD 2 ]+---------------------
> slot_name    | replica_local_slot
> plugin       |
> slot_type    | physical
> datoid       |
> database     |
> active       | t
> active_pid   | 2043
> xmin         | 5463120
> catalog_xmin |
> restart_lsn  | 9/11FA3D18

Ídem; además, el xmin es 8 millones que no está ni cerca del
Latest checkpoint's NextXID:          0/2021183943
(dos mil millones) que mostraste en controldata.

Wtf?

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

-
Enviado a la lista de correo pgsql-es-ayuda (pgsql-es-ayuda@postgresql.org)
Para cambiar tu suscripción:
http://www.postgresql.org/mailpref/pgsql-es-ayuda

Responder a