> As I understand it, pgpool-ii is set up so that a backend of servers sit > and wait for communications from the client-facing pgpool-ii server. > Since pgpool acts as statement based replication, is it possible to set > up more than two pgpool-ii servers facing clients? > > I am curious as this would be a method of load-balancing even further. > For example, if pgpool-ii server has a high load on it, it would make > sense to have another pgpool server serving requests, even if they're > pointing to the same backend servers. It would offer another level of > load balancing capability. > > Example in diagram form: > > level3. backend1, backend2, backend3 <----- backend postgres servers > level2. pgpool1, pgpool2 <---- pgpool servers working with backend1, 2 > and 3 > level1. web servers A,B,C,D,E,F <------ Clients of the pgpool servers > > Is this possible? Am I right in my belief that the backend postgres > servers themselves do not care or know if they're in a pool, that is, > the only bit of software that makes this logically into a pool of > servers is pgpool itself?
I believe the configuration above works. -- Tatsuo Ishii SRA OSS, Inc. Japan _______________________________________________ Pgpool-general mailing list [email protected] http://pgfoundry.org/mailman/listinfo/pgpool-general
