Skip site navigation (1) Skip section navigation (2)

Re: Postgres 9.1 Synchronous Replication and stuck queries during sync repl setup

From: Manoj Govindassamy <manoj(at)nimblestorage(dot)com>
To: Gabriele Bartolini <Gabriele(dot)Bartolini(at)2ndQuadrant(dot)it>
Cc: "pgsql-admin(at)postgresql(dot)org" <pgsql-admin(at)postgresql(dot)org>
Subject: Re: Postgres 9.1 Synchronous Replication and stuck queries during sync repl setup
Date: 2012-06-06 16:16:16
Message-ID: ADD354851455FC44A5D8923AD5152DE307A0FA90@coloex01.nimblestorage.com (view raw or flat)
Thread:
Lists: pgsql-adminpgsql-general
Gabriele,

Thanks for the quick reply.

> That's not necessary. Usually you do this only the first time you set
> it up, then take advantage of the wal_keep_segments on the master and
> allow the standby to resync.

we are not sure about when the standby will come again and sync with master. also, we have space constraints on the master and can't keep indefinite segments.

> This behaviour is perfectly fine. Until the master and the standby are
> in sync,

But, even after the standby came to sync mode, those statements that are executed between (1) and (2) remain stuck only. they don't complete at all.

> Then uncomment the 'synchronous_standby_names' line on the
> master and issue a reload.

doesn't any change to 'synchronous_standby_names' need a server restart ?? Just a reload config command is sufficient here ?

yes, all the servers are in same LAN. Thanks for your time.


- manoj

________________________________________
From: pgsql-admin-owner(at)postgresql(dot)org [pgsql-admin-owner(at)postgresql(dot)org] on behalf of Gabriele Bartolini [Gabriele(dot)Bartolini(at)2ndQuadrant(dot)it]
Sent: Wednesday, June 06, 2012 8:51 AM
To: Manoj Govindassamy
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: [ADMIN] Postgres 9.1 Synchronous Replication and stuck queries during sync repl setup

 Hi,

 On Wed, 6 Jun 2012 14:55:15 +0000, Manoj Govindassamy
 <manoj(at)nimblestorage(dot)com> wrote:
> PG Slave gets fresh backup from PG master using pg_backup utility
> everytime before it starts up

 That's not necessary. Usually you do this only the first time you set
 it up, then take advantage of the wal_keep_segments on the master and
 allow the standby to resync.

> A. I need to know why PG master started accepting connections at (1)
> and still NOT able to fully commit the transactions. Statements that
> are executed after (3) are not seeing this problem.

 This behaviour is perfectly fine. Until the master and the standby are
 in sync, given that you set the standby to be the synchronous one, the
 master MUST wait until the standby writes on disk the transaction
 information.

 I suggest that first you remove the standby server from the list of
 synchronous servers by commenting the 'synchronous_standby_names' line
 and wait until the standby catches up (asynchronous streaming
 replication). Then uncomment the 'synchronous_standby_names' line on the
 master and issue a reload. From that moment on you will have synchronous
 streaming replication in place.

 Cheers,
 Gabriele

 P.S.: I took it for granted that the two servers are in the same LAN.
--
  Gabriele Bartolini - 2ndQuadrant Italia
  PostgreSQL Training, Services and Support
  Gabriele(dot)Bartolini(at)2ndQuadrant(dot)it - www.2ndQuadrant.it

--
Sent via pgsql-admin mailing list (pgsql-admin(at)postgresql(dot)org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin



In response to

pgsql-admin by date

Next:From: Tom LaneDate: 2012-06-06 16:24:35
Subject: Re: postgres block size alignment with filesystem block size
Previous:From: Mike BroersDate: 2012-06-06 15:54:40
Subject: postgres block size alignment with filesystem block size

pgsql-general by date

Next:From: Benson JinDate: 2012-06-06 16:20:06
Subject: Need help in transferring FP to Int64 DateTime
Previous:From: Frank LanitzDate: 2012-06-06 15:58:28
Subject: Re: pg_database_size differs from df -s

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group