Re: Re: xReader, double-effort (was: Temporary tables under hot standby)

From: Aakash Goel <aakash(dot)bits(at)gmail(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Kevin Grittner <Kevin(dot)Grittner(at)wicourts(dot)gov>, simon(at)2ndquadrant(dot)com, andres(at)2ndquadrant(dot)com, josh(at)postgresql(dot)org, pgsql-hackers(at)postgresql(dot)org
Subject: Re: Re: xReader, double-effort (was: Temporary tables under hot standby)
Date: 2012-04-28 16:08:49
Message-ID: CAAEmBAC8x9WYs-cpnSU8NJCfU1duJgMk7x_qYNHuCjbwTS=AYA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hello Tom,

> I'm not sure I
> get the point of logical replication that requires a physical replica as
> a prerequisite.

> It would be enormously
> more performant for the master to be emitting logical replication
> records to start with, since it already has all the right names etc
> at hand at basically no cost.

I want to emphasize that any system which uses the logical reader setup,
including a replication system, tends to be very computationally intensive
on the database which it queries all so often. In most of the environments,
the source database is the main database, and as such, any performance
degradation on this database is very bad.

On the other hand, if we offload almost all the work to the physical
replica, our source database, which is the main database, still functions
at the same throughput.

Thus, at the cost of having the replication system as a whole run a little
slower than it could if it were using the main database, we have made sure
that our performance critical main source database is not affected in
performance at all.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Aakash Goel 2012-04-28 16:31:46 Re: Welcome 2012 GSOC students
Previous Message Tom Lane 2012-04-28 15:43:09 Re: Re: xReader, double-effort (was: Temporary tables under hot standby)