Re: Patch for fail-back without fresh backup

From: Pavan Deolasee <pavan(dot)deolasee(at)gmail(dot)com>
To: Sawada Masahiko <sawada(dot)mshk(at)gmail(dot)com>
Cc: Simon Riggs <simon(at)2ndquadrant(dot)com>, Samrat Revagade <revagade(dot)samrat(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Patch for fail-back without fresh backup
Date: 2013-06-25 03:19:10
Message-ID: CABOikdM7x=p0ymKNrVV6P+9Z77UpzkraV7F_4GSu2CDXQHUW+w@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Jun 24, 2013 at 7:17 PM, Sawada Masahiko <sawada(dot)mshk(at)gmail(dot)com>wrote:

>
>
> --------------------
> [Server]
> standby_name = 'slave1'
> synchronous_transfer = commit
> wal_sender_timeout = 30
> [Server]
> standby_name = 'slave2'
> synchronous_transfer = all
> wal_sender_timeout = 50
> -------------------
>

What different values/modes you are thinking for synchronous_transfer ?
IMHO only "commit" and "all" may not be enough. As I suggested upthread, we
may need an additional mode, say "data", which will ensure synchronous WAL
transfer before making any file system changes. We need this separate mode
because the failback safe (or whatever we call it) standby need not wait on
the commits and it's important to avoid that wait since it comes in a
direct path of client transactions.

If we are doing it, I wonder if an additional mode "none" also makes sense
so that users can also control asynchronous standbys via the same mechanism.

Thanks,
Pavan

--
Pavan Deolasee
http://www.linkedin.com/in/pavandeolasee

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2013-06-25 03:56:50 Re: [9.4 CF 1] The Commitfest Slacker List
Previous Message Steve Singer 2013-06-25 03:16:14 Re: [PATCH] Fix conversion for Decimal arguments in plpython functions