Message-ID: <4ED44582.4080306@ringerc.id.au>
Date: Tue, 29 Nov 2011 10:37:54 +0800
From: Craig Ringer <ringerc@ringerc.id.au>
User-Agent: Mozilla/5.0 (X11; Linux x86_64;
	rv:7.0.1) Gecko/20110930 Thunderbird/7.0.1
MIME-Version: 1.0
To: Lukasz Brodziak <lukasz.brodziak@gmail.com>
CC: pgsql-admin@postgresql.org
Subject: Re: Sequence "self" change - possible?
References: 
 <CAGWYGjVR6bwNpAwFRF-_NRP1fwFZCDWspyxLH58y--uW7ZPiHA@mail.gmail.com>
In-Reply-To: 
 <CAGWYGjVR6bwNpAwFRF-_NRP1fwFZCDWspyxLH58y--uW7ZPiHA@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 11/28/2011 10:04 PM, Lukasz Brodziak wrote:
> Hello,
>
> Is it possible for sequence to decrement its value by itself? Or is it
> possible to see wether it was changed by someone. The thing is that on
> one of our clinet's DB sequence rolled back to the state from 8 months ago.

It shouldn't generally be possible, no. It'd have to be changed by a 
call to nextval(), setval() or the like, or a direct UPDATE to the 
underlying sequence pseudo-table.

Could someone have tried to restore a backup from 8 months ago? That'd 
result in a setval(...) call when pg_restore or psql (depending on 
backup format) tried to set the sequence position.

In your case, though, your next post mentions possible database 
corruption, so I'm wondering if this is related. If something's been 
scribbling over the database files, someone's half-restored an old 
file-level backup, or something similarly horrible then anything could 
happen.

--
Craig Ringer