...

> Could you provide a complete test case?
As I've been told, there is no special code, just a usual sequence usage. There is no small part of code that can be
reproduced.
--
The situation is that there are 2 applications working with the same DB and accessing sequences. Each time application
isstarted the sequences are reloaded (without doing this the sequences go crazy, e.g. you can get 1000 with 5 being the
lastvalue). 
--
> Is it possible that another session is altering the sequence to
> start with a lower value?
No, as only next value is called.

Have you perchance set the sequence's
> CACHE setting to a value other than 1 (one)?
No.

> What's the output of
> "SELECT * FROM sequence_name"?
ERROR:  relation "sequence_name" does not exist

> Are you making queries from more
> than one session?  Are you using connection pooling?
Yes. No.

Thank you.


On 4/7/06, Michael Fuhr <mike@fuhr.org> wrote:
> On Thu, Apr 06, 2006 at 10:04:03AM +0000, Alex Fomin wrote:
> > While using the following function:
> > ---
> > nextval(sequence_name)  returns currval(sequence_name) -1
> > ---
> > while +1 is expected. It happens only sometimes, no dependency can be found.
>
> Could you provide a complete test case?  That is, all SQL statements
> that somebody could execute in an empty database to reproduce the
> problem.  It doesn't have to be 100% reproducible as long as it
> does exhibit the behavior every once in a while (an indication of
> how often or under what circumstances, if known, would be helpful).
>
> Is it possible that another session is altering the sequence to
> start with a lower value?  Have you perchance set the sequence's
> CACHE setting to a value other than 1 (one)?  What's the output of
> "SELECT * FROM sequence_name"?  Are you making queries from more
> than one session?  Are you using connection pooling?
>
> --
> Michael Fuhr
>
>

--


Bis zu 70% Ihrer Onlinekosten sparen: GMX SmartSurfer!
      Kostenlos downloaden: http://www.gmx.net/de/go/smartsurfer