[Pgpool-general] Confusing... pgpool starts on both nodes, but don't replicate.

[J. Carlos Muro]

As Gerd already pointed, you need to activate just one pgpool, then connect
to it from your sql client. (I leave here the question for experts: what
happens if I set up two or more pgpool all of them accessing to the same
backends at the same time?)

As a guideline to find the problem to your setup, you can enable debug mode
and try to find kind of these lines in your logs:

[....]
2009-06-30 16:29:15 DEBUG: pid 11297: backend 0 weight: 1073741823.500000
2009-06-30 16:29:15 DEBUG: pid 11297: backend 1 weight: 1073741823.500000
[....]
2009-06-30 16:29:15 LOG:   pid 11297: pgpool successfully started

>From then an on, when you try to connect pgpool in order to execute queries
you can check netstat as to see if pgpool child opened connection to both of
your backends:

$ sudo netstat -puta | grep pool
[....]
tcp        0      0 pgpoolserver:51190          backend0:postgresql
ESTABLISHED 11319/pgpool
tcp        0      0 pgpoolserver:36816          backend1:postgresql
ESTABLISHED 11319/pgpool

Also in your logs you should find something like this:

2009-06-30 16:30:47 DEBUG: pid 11319: new_connection: connecting 0 backend
2009-06-30 16:30:47 DEBUG: pid 11319: new_connection: connecting 1 backend
[....]
2009-06-30 16:30:47 DEBUG: pid 11319: select_load_balancing_node: selected
backend id is 0
2009-06-30 16:30:47 DEBUG: pid 11319: read_kind_from_backend: read kind from
0 th backend Z NUM_BACKENDS: 2
2009-06-30 16:30:47 DEBUG: pid 11319: read_kind_from_backend: read kind from
1 th backend Z NUM_BACKENDS: 2

Hope it helps!
J. Carlos Muro

2009/7/2 » Ruben Alves <rubenalves at fm.ul.pt>

> Hi everyone,
> I installed pgpool with postgreSQL on Monday. I've been busy since
> yesterday morning, and in the afternoon I've tried to configure following
> those instructions:
>
>   *
> http://pgpool.projects.postgresql.org/pgpool-II/doc/pgpool-en.html#start
>   *
> http://pgpool.projects.postgresql.org/pgpool-II/doc/tutorial-en.html#replication
>
> Everything is running fine, pgptool is running, listening for news
> connections etc.. But NOTHING happens. If I change something in node1
> (docsys) nothing is replicated on node2 (docsys2).
>
> I have this king of configuration (exactly the same file for node1 and
> node2):
>
> [....]
> # Replication mode
> replication_mode = true
>
> # Load balancing mode, i.e., all SELECTs are load balanced.
> load_balance_mode = true
>
> # if there's a data mismatch between master and secondary
> replication_stop_on_mismatch = false
> [...]
>
> [...]
> #set pgpool2 hostname
> pgpool2_hostname = 'docsys'
>
> # system DB info
> system_db_hostname = 'docsys'
> system_db_port = 5432
> system_db_dbname = 'pgbench'
> system_db_schema = 'pgbench_public'
> system_db_user = 'pgpool_001'
> system_db_password = 'pass'
>
> # backend_hostname, backend_port, backend_weight
> # here are examples
> backend_hostname0 = 'docsys'
> backend_port0 = 5432
> backend_weight0 = 1
> backend_data_directory0 = '/backup/postgres/data_001/'
>
> backend_hostname1 = 'docsys2'
> backend_port1 = 5433
> backend_weight1 = 1
> backend_data_directory1 = '/backup/postgres/data_002/'
> [...]
>
>
> (I've just paste parts of the config file in order to not flood this ML)
>
> I don't really understand how pgpool initialize the replication system.
> I've build a MySQL replication system (master-slaves) and maybe i'm too
> focused on this past experience.
>
> Can somebody give me a hand ?
>
>
>
>
> --
> Ruben Alves
> System Analyst USI @ Instituto de Medicina Molecular
>
> _______________________________________________
> Pgpool-general mailing list
> Pgpool-general at pgfoundry.org
> http://pgfoundry.org/mailman/listinfo/pgpool-general
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pgfoundry.org/pipermail/pgpool-general/attachments/20090702/895efbbb/attachment.html>