Re: Replication with a third-party plug-in

From: bricklen <bricklen(at)gmail(dot)com>
To: Craig James <cjames(at)emolecules(dot)com>
Cc: "pgsql-admin(at)postgresql(dot)org" <pgsql-admin(at)postgresql(dot)org>
Subject: Re: Replication with a third-party plug-in
Date: 2017-11-17 16:02:00
Message-ID: CAGrpgQ9np208vCAgaowSKmDchfb0NEAH9-_Nsn31CaOQT9otCg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

On Fri, Nov 17, 2017 at 7:53 AM, Craig James <cjames(at)emolecules(dot)com> wrote:

> The problem is with replication; it's a sort of chicken-and-egg problem.
> Installation of Jchem-psql requires some transactions to happen, but on a
> standby server, transactions aren't allowed. So you can't install
> Jchem-psql on the slave server. That means that a hot-standby server
> doesn't work, because even if the Postgres database is replicated, chemical
> queries can't be executed. Worse, if I need to do a failover to the standby
> server, I have to install Jchem-psql after the failover, then recreate all
> of the chemical indexes, which takes many hours.
>

An alternate solution could potentially be to use logical replication if
you are on 9.4+. The "replica", which is in fact another master database
getting changes fed to it by the logical replication process, can have all
the changes from the master fed to it in the form of queries which will
trigger your Jchem-psql tool.

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Don Seiler 2017-11-17 16:05:14 Re: FATAL: invalid value for parameter "TimeZone" after upgrade from 9.2 to 9.6
Previous Message Tom Lane 2017-11-17 15:57:26 Re: FATAL: invalid value for parameter "TimeZone" after upgrade from 9.2 to 9.6