[pgpool-general: 586] Combining replication flavors

helpdesk at acha.nl helpdesk at acha.nl
Mon Jun 4 19:53:08 JST 2012


Hi,

 

We have been successfully using pg-pool for some time with a master/master setup. We are now looking for a way to combine master/master replication with master/slave replication in one setup. 

Is that possible with pg-pool?

 

Use case scenario:

We need a production cluster (4db's - master/master) and a report cluster (also 4 db's). The idea is to send heavy report requests to the report cluster in order to relieve the production cluster. It would be nice if the report cluster is quite recent, but it is not a problem if it falls a bit behind.

 

I have been thinking of a solution where one of the master/master node is another pg-pool master which takes care of the master/slave replication using PostgreSQL's streaming replication for example. But that setup has some down sides: it introduces a single point of failure (or we need to add more complexity for a redundant setup) and I assume the production pg-pool master has to wait for the report pg-pool master to finish it's write action. Which may take a while when it is busy processing report data.

 

Any idea's are very much appreciated!

 

Thanks guys,

Marco


--
This message was scanned by ESVA and is believed to be clean.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20120604/e08c6325/attachment.html>


More information about the pgpool-general mailing list