Replication slots for cascading replication.

From: Tore Halvorsen <tore(dot)halvorsen(at)gmail(dot)com>
To: "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org>
Subject: Replication slots for cascading replication.
Date: 2018-02-05 08:20:57
Message-ID: CADGw-SeqhCzKGYeDEeUPXs-ccXKB8VPJALh369Nej1-YBaV5Uw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Hi,

I encountered an issue this weeked and was wondering if there is a simple
solution to avoid this specific problem.

A simplified database setup here i three servers, which uses streaming
replication with replication slot on pg10.1
M: a "master" server
C: a hot standby that cascades to ...
S: a slave.

So data data flow is
M -> C -> S.

There is a replication slot for C on M - and a slot for S on C.

S was taken down for mainenance - and during this window C suffered a
hardware failure which redered the machine dead.

Switching S to restore from M instead of C lead to missing WAL files, since
the lsn for the C slot on M had already passed to lsn on S.

So - is there a way to pass the cascaded info upstream? E.g. is there a way
to non-manually make sure the WAL files on M are present for both C and S?

(Hopefully this understandable)

Browse pgsql-general by date

  From Date Subject
Next Message Thomas Poty 2018-02-05 10:26:33 weird result by changing type enum array to text array of a column
Previous Message Khushboo Vashi 2018-02-05 03:53:36 Re: pgAdmin questions