BUG #5476: sequence corruption - Mailing list pgsql-bugs

From Piergiorgio Buongiovanni
Subject BUG #5476: sequence corruption
Date
Msg-id 201005270910.o4R9AKDd038676@wwwmaster.postgresql.org
Whole thread Raw
List pgsql-bugs
The following bug has been logged online:

Bug reference:      5476
Logged by:          Piergiorgio Buongiovanni
Email address:      p.buongiovanni@net-international.com
PostgreSQL version: 8.3.4
Operating system:   CentOS (Redhat 5.4)
Description:        sequence corruption
Details:

We experimented a sequence corruption on a table:
If we now look at the sequence data using pgAdmin version 1.8.4, we see the
following:

CREATE SEQUENCE business.subject_isid_seq
  INCREMENT 1
  MINVALUE 1
  MAXVALUE 9223372036854775807
  START 58827944
  CACHE 1;
ALTER TABLE business.subject_isid_seq OWNER TO netdw_owner;

The problem is that now the search on the table business.Subject (where the
sequence is used) performs very slowly. It seems the data are corrupted but
at the end I can see that there is a record created with the number 463663
instead of 57. As you can see the next value will be used is 58827944.
Which is the problem? Can I recover this situation and restore the right
sequence value? I tried to set the START value to 58 but, as you can see,
the value changed.

pgsql-bugs by date:

Previous
From: Jasen Betts
Date:
Subject: Re: BUG #5469: regexp_matches() has poor behaviour and more poor documentation
Next
From: Boszormenyi Zoltan
Date:
Subject: Re: Cache lookup failure for index during pg_dump