Re: Continuent (was: Postgres Clustering)

From: Flavio Henrique Araque Gurgel
Subject: Re: Continuent (was: Postgres Clustering)
Date: ,
Msg-id: 15890762.21243556788315.JavaMail.flavio@presente
(view: Whole thread, Raw)
In response to: Continuent (was: Postgres Clustering)  (Alan McKay)
List: pgsql-performance

----- "Alan McKay" <> escreveu:
> Hmmm.   Anyone out there have the Continuent solution working with PostgreSQL?
> If so, what release?  We're at 8.3 right now.

I have tested Sequoia 2.10.10 with a high transaction rate database with good servers and plenty of memory. Since that's a OLTP system the performance droped to as low as 30%.

I can't recomend their solution for a OLAP system because I never tested in this situation.

Configuration of Sequoia is quite complicated and I think a very good Database Administrator is needed to keep it working correctly and nodes syncronized.
Sequoia also is very complicated to run ddl and dml scripts since your scrips should be written for Sequoia, not for PostgreSQL.

If log-shipping works for you, try Slony. Your slaves can serve as read-only databases and you can distribute some load.

Flavio

pgsql-performance by date:

From: Greg Smith
Date:
Subject: Re: Scalability in postgres
From: Grzegorz Jaśkiewicz
Date:
Subject: Re: Scalability in postgres