Re: PostgreSQL clustering with DRBD
От | Scott Marlowe |
---|---|
Тема | Re: PostgreSQL clustering with DRBD |
Дата | |
Msg-id | dcc563d10903021822j25e8b667qb76993c11eef39dd@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: PostgreSQL clustering with DRBD (Serge Fonville <serge.fonville@gmail.com>) |
Ответы |
Re: PostgreSQL clustering with DRBD
|
Список | pgsql-general |
On Mon, Feb 23, 2009 at 2:27 AM, Serge Fonville <serge.fonville@gmail.com> wrote: > Thanks all for the responses, >> >> We're very happy with pgpool-II for load-balancing and multi-master >> >> usage of PostgreSQL (keep in mind to enable HA for pgpool-II itself to >> >> avoid a SPOF, e.g. with heartbeat). > > I could not determine whether pgpool-II is suitable for what I want. > It does not seem to support multimaster in the fashion I had in mind, based > on the information on the website it looks like it does not support full > CRUD on any node. So, what design criteria are you using that says multi-master is a better choice than master / slave? I know that multi-master is buzzword compliant, but often the actual product you get with it isn't any better, and in some cases worse, than a master / slave setup. > The most suitable solution seems to be LVS for a shared IP, ldirectord for > load balancing and cybercluster for the database > replication/synchronization. Again, this is a lot of work to avoid master / slave with failover. Are you sure it's really needed for your situation?
В списке pgsql-general по дате отправления: