El Thu, Apr 04, 2002 at 08:58:50AM +0800, Patrick Hsieh escribió: > Hello Jean-Francois Dive <jef@linuxbe.org>, > > I have no problem with the robustness of PostgresSQL. > However, does PostgreSQL have built-in replication? If not, how do I > implement high-available PostgreSQL clusters? Umm, AFAIK, 7.X series of PostgreSQL have built-in support for table replication, check de features page at their site to confirm. No matter if native supported or not .. there are some *load balancing* daemons around Inet, that let you setup a group of psql servers. Time ago I worked with one of that daemons and did this config: Node A: Node B: Node C: Node D: psql psql psql psql Daemon setup: Daemon Setup: Daemon Setup: Daemon Setup: Rep. to B Rep. to A Rep. to A Rep. to A Rep. to C Rep. to C Rep. to B Rep. to B Rep. to D Rep. to D Rep. to D Rep. to C It worked, but it's a big problem to have them perfecly in sync, because if you use RR at the dns level to do another *load balancing* task, uff, get harder to sync, beleive me. The best, if you target is HA and not Load Balancing is this one the same as above, but not doing DNS RR, instead, DNS failover, if master fails, slave get on running. Easier to keep in sync, because all the sql statements goes from Master to Slaves, if master goes down, one of the slaves becomes master and all contine running. In this aproach, a *faster* solution is to do IP takeover, if master an slaves are on the same subnet. Best Regards -- _ _ // Raúl A. Betancort Santana /> A Dream is an answer to __ \\ // <rabs@dimension-virtual.com> // question that we don't know (oo) \\ // Dimensión Virtual S.L. // how to ask. / \/ \ // \> A Linux Solution Provider </ `V__V' </
Attachment:
pgpcr0B2_Fk90.pgp
Description: PGP signature