Re: Fully-automatic streaming replication failover when master dies?

From: Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>
To: Dmitry Koterov <dmitry(dot)koterov(at)gmail(dot)com>
Cc: Postgres General <pgsql-general(at)postgresql(dot)org>
Subject: Re: Fully-automatic streaming replication failover when master dies?
Date: 2014-01-24 02:35:21
Message-ID: CAOR=d=0knRgqXFt9Yw7hEEJtxwdcK=P66RjHQDTLaivONJmBQQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Currnently there is no automatic failover for postgresql. If I were
to build something to do this I'd start with repmgr:
http://www.repmgr.org/

On Wed, Jan 22, 2014 at 2:35 PM, Dmitry Koterov
<dmitry(dot)koterov(at)gmail(dot)com> wrote:
> Hello.
>
> I googled 1 hour approximately, but have not found a ready solution for
> this. So maybe this feature is in PostgreSQL todo-list, or something similar
> exists somewhere...
>
> Before the actual question, I'd like to give a small analogy. What I mostly
> love in MongoDB is that it supports a fully transparent scheme of
> replication failover. If you have >= 3 MongoDB notes (e.g. 1 master and 2
> replicas), and the master dies, in a couple of seconds a replica is
> AUTOMATICALLY elected as a new master, and all other replicas are
> AUTOMATICALLY begin to follow it. If the dead master is back again suddenly,
> it first appears as a replica, but in a couple of seconds it becomes a new
> master back (because it initially had a highest weight assigned), and all
> replicas become to follow it. All these steps are done automatically and
> transparently. It just works.
>
> So does something similar and more-or-less stable exist for PostgrSQL too?
>
> P.S.
> The links I've found already:
> - http://www.databasesoup.com - that's it, but too young and, possibly, not
> yet stable
> - repmgr: always referred at StackOverflow (thousands of times), but it does
> not provide an automatic failover :-)
> - pgpool-2: it contains a couple of config options for "failover", but in
> practice they are too poor to be used in production and with large
> databases, I believe after reading the docs
> - pacemaker's plugin: extremely complex (IMHO) overweighted for this purpose
> - EnterpriseDB's solutions: they are not free

--
To understand recursion, one must first understand recursion.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Scott Marlowe 2014-01-24 02:41:54 Re: Fully-automatic streaming replication failover when master dies?
Previous Message Sameer Kumar 2014-01-24 02:18:27 Re: problem connecting to postgres via apache