Re: I'd like to discuss scaleout at PGCon

On Sun, Jun 3, 2018 at 2:00 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
>
> In XL, GTM is a singe component managing transaction ids. That has a
> standby, so is not a SPOF.
>
> But that is not what I mean. I don't believe that a GTM-style
> component is necessary in a future in-core scalablility solution.
>

I agree. I think, a distributed algorithm, which does not need a
single GTM-style node, would be better. That automatically provides
high availability without configuring a standby.

> Each incoming query needs to be planned and executed from one
> coordinator component, then the work performed across many workers on
> different nodes (or just one).

Each node need to be confiugred and maintained. That requires efforts.
So we need to keep the number of nodes to a minimum. With a
coordinator and worker node segregation, we require at least two nodes
in a cluster and just that configuration doesn't provide much
scalability. With each node functioning as coordinator (facing
clients) and worker (facing other coordinators) keeps the number of
nodes to a minimum. It is good for HA.

--
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company