[pgpool-general: 280] Is there a reason for PostgreSQL 9.1/Pgpool 3.1?

Lumir Jasiok lumir.jasiok at vsb.cz
Thu Mar 15 00:17:53 JST 2012


Hi,

I am currently working on PG HA server solution for our university with 
following schema:


  client1....clientn
          |
          |
          |  9999
+-----------+
  | pgpool         |
+-----------+
              \
                \
                  \---\
                           \
+---------+      \            +---------+
  | pgpool1    |--heartbeat--|pgpool2    |
+---------+                    +---------+
       |                                             |
      +------------------------+
                     /  \
                    /    \
                   /      \
                  /        \
                 /          \
                /            \
+-----------+     +-----------+
  |  pg1             |       |   pg2            |
+-----------+     +-----------+

pgpool - frontend IP address (heartbeated)
pgpool1 - pgpool server 1
pgpool2 - pgpool server 2
pg1 - PostgreSQL backend server 1
pg2 - PostgreSQL backend server 2

We are using Ubuntu 10.04 with PostgreSQL 8.4 and Pgpool-II 2.3.

We want to use pgpool-II for replication and load balancing. Is there a 
good technical reason (some feature) for using PostgreSQL 9.1 and/or 
Pgpool-II 3.1?

Can we trust (online recovery process, WAL, PITR), that the data will be 
always synced?

Is this solution - heartbeated pgpool servers, two and more PostgreSQL 
backends common/preferred? Do you have some advices, tips...?

Best Regards

Lumir

-- 
  Lumír Jasiok
  VSB-TU Ostrava - Computer centre



More information about the pgpool-general mailing list