Re: Pgpool questions - Mailing list pgsql-general

From Tatsuo Ishii
Subject Re: Pgpool questions
Date
Msg-id 20050427.114226.71088639.t-ishii@sra.co.jp
Whole thread Raw
In response to Re: Pgpool questions  ("Brian Maguire" <bmaguire@vantage.com>)
List pgsql-general
Ok, I have added per query basis weight control to my TODO list.
--
Tatsuo Ishii

> Thanks.
>
>
>
> Is it on your todo list to something like this...?  It is the opposite
> of "/* NO LOAD BALANCE */".
>
>
>
> Make a "/* SLAVE */" type comment to force it to a slave? It would be
> helpful in forcing a query to a reporting server if you know your what
> your reporting queries are.
>
>
>
> "Set weight_secondary to 0."
>
>
>
> Did you mean Set weight_master to 0.  Unfortunately this would not help
> on a query by query basis, but the above would.
>
>
>
>
>
>
>
>
>
>
>
>
>
> -----Original Message-----
> From: Tatsuo Ishii [mailto:t-ishii@sra.co.jp]
> Sent: Saturday, April 23, 2005 7:23 PM
> To: Brian Maguire
> Cc: pgsql-general@postgresql.org
> Subject: Re: [GENERAL] Pgpool questions
>
>
>
> > I have a couple Pgpool questions regarding running pgpool as a load
> balancer with Sony as the replicator.
>
> >
>
> >
>
> >
>
> > Is there anyway to load balance selects to more than 2 servers? ex.  1
> master with 2 slaves?  Run multiple instances?
>
>
>
> It's on my TODO but will not be a trivial change.
>
>
>
> > Is there any way to tell a select (ex. with a function) to go only the
> master and not the slave?
>
>
>
> Pgpool will send any query not starting with "SELECT" or "select" to
>
> the master. So you can insert a white space or a comment, for example
>
>  "/* NO LOAD BALANCE */" at the beginning of the line to prevent the
>
>  query to be sent to the slave.
>
>
>
> > Is there any way to tell a select query to only go to the slave?
>
>
>
> Set weight_secondary to 0.
>
> --
>
>
>
> Tatsuo Ishii
>

pgsql-general by date:

Previous
From: James Robinson
Date:
Subject: Re: Performance difference between ANY and IN, also array syntax
Next
From: Tom Lane
Date:
Subject: Re: subqueries and qualification of table names