[Pgpool-general] Simple pgpool II setup, strange error

Fri May 22 00:27:34 UTC 2009

>    I'm testing a scenario where I will use PgPool II and Slony I as
> replication and HA solution.
>    Tested in one machine with two virtual servers, using Postgres
> 8.3.7 ans PgPool II 2.2.2.
>    Worked flawlessly, less the PITR recovery, as said on the documentation.
> 
>    Now, on the same machine, using two different ports and $PGDATA
> directories, after startup,
> PgPool is ok.
>    If I connect using psql or pgadmin, and run commands like "select
> 1;", I'm consistently getting this log errors.
> 
>   2009-05-20 12:22:06 LOG:   pid 10299: pgpool successfully started
> 2009-05-20 12:22:06 DEBUG: pid 10332: I am PCP 10332
> 2009-05-20 12:22:06 DEBUG: pid 10299: starting health checking
> 2009-05-20 12:22:06 DEBUG: pid 10299: health_check: 0 th DB node status: 1
> 2009-05-20 12:22:06 DEBUG: pid 10299: health_check: 1 th DB node status: 1
> 2009-05-20 12:22:14 DEBUG: pid 10303: I am 10303 accept fd 5
> 2009-05-20 12:22:14 DEBUG: pid 10303: Protocol Major: 3 Minor: 0
> database: postgres user: postgres
> 2009-05-20 12:22:14 DEBUG: pid 10303: new_connection: connecting 0 backend
> 2009-05-20 12:22:14 DEBUG: pid 10303: new_connection: connecting 1 backend
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length: slot:
> 0 length: 8
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length: slot:
> 1 length: 8
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name:
> client_encoding value: UTF8
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name:
> client_encoding value: UTF8
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 23
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 23
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name: DateStyle
> value: ISO, MDY
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name: DateStyle
> value: ISO, MDY
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 26
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 26
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name:
> integer_datetimes value: off
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name:
> integer_datetimes value: off
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 20
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 20
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name: is_superuser value: on
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name: is_superuser value: on
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name:
> server_encoding value: UTF8
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name:
> server_encoding value: UTF8
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name:
> server_version value: 8.3.7
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name:
> server_version value: 8.3.7
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 35
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 35
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name:
> session_authorization value: postgres
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name:
> session_authorization value: postgres
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 36
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 36
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name:
> standard_conforming_strings value: off
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name:
> standard_conforming_strings value: off
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 0 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length2:
> master slot: 1 length: 25
> 2009-05-20 12:22:14 DEBUG: pid 10303: 0 th backend: name: TimeZone
> value: Brazil/East
> 2009-05-20 12:22:14 DEBUG: pid 10303: 1 th backend: name: TimeZone
> value: Brazil/East
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length: slot:
> 0 length: 12
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length: slot:
> 1 length: 12
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_send_auth_ok: send pid
> 10340 to frontend
> 2009-05-20 12:22:14 DEBUG: pid 10303: select_load_balancing_node:
> selected backend id is 0
> 2009-05-20 12:22:14 DEBUG: pid 10303: read_kind_from_backend: read
> kind from 0 th backend Z NUM_BACKENDS: 2
> 2009-05-20 12:22:14 DEBUG: pid 10303: read_kind_from_backend: read
> kind from 1 th backend Z NUM_BACKENDS: 2
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_process_query: kind from backend: Z
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length: slot:
> 0 length: 5
> 2009-05-20 12:22:14 DEBUG: pid 10303: pool_read_message_length: slot:
> 1 length: 5
> 2009-05-20 12:22:14 DEBUG: pid 10303: ReadyForQuery: message length: 5
> 2009-05-20 12:22:14 DEBUG: pid 10303: ReadyForQuery: transaction state: I
> 2009-05-20 12:22:16 DEBUG: pid 10299: starting health checking
> 2009-05-20 12:22:16 DEBUG: pid 10299: health_check: 0 th DB node status: 2
> 2009-05-20 12:22:16 DEBUG: pid 10299: health_check: 1 th DB node status: 2
> 2009-05-20 12:22:20 DEBUG: pid 10303: read kind from frontend X(58)
> 2009-05-20 12:22:26 DEBUG: pid 10299: starting health checking
> 
>    The status change to 2, although the replication continues without a hitch.
>    After restarting, the status goes back to 1.
>    Is there a workaround to this problem ?

Did you see any errors or notices in PostgreSQL log around 12:22:16
when the health checking failed? 
--
Tatsuo Ishii
SRA OSS, Inc. Japan