Re: [HACKERS] two records with same oid, freak accident? - Mailing list pgsql-hackers

From Massimo Dal Zotto
Subject Re: [HACKERS] two records with same oid, freak accident?
Date
Msg-id 199807181532.RAA07614@pennac.cs.unitn.it
Whole thread Raw
In response to two records with same oid, freak accident?  (Brett McCormick <brett@work.chicken.org>)
List pgsql-hackers
>
>
> in some sort of freak accident, we've ended up with a duplicated
> record.  all info, including the oid was duplicated.
>
> it was during an update, two people ran a command at the same time.
>
> user 1: BEGIN
> user 1: NOTIFY
> user 1: UPDATE
> user 2: BEGIN
> user 2: NOTIFY
> user 1: END
> user 2: UPDATE
> user 2: END
>
> same command, so the queries are the same.  the record duplicated was
> the one being updated.  i'll try to reproduce it.
>
> also, I had a unique index on the table, but that didn't seem to make
> any difference.
>
> any ideas on how to delete one without deleting both?
>

I have the same problem with pg_listeners. Sometimes I find duplicate records
with same oid in the table inserted by concurrent transactions. I suspect
that the problem is caused by the notify but I'm not sure. Could you post
some test commnds to reproduce the problem ?

--
Massimo Dal Zotto

+----------------------------------------------------------------------+
|  Massimo Dal Zotto                e-mail:  dz@cs.unitn.it            |
|  Via Marconi, 141                 phone:  ++39-461-534251            |
|  38057 Pergine Valsugana (TN)     www:  http://www.cs.unitn.it/~dz/  |
|  Italy                            pgp:  finger dz@tango.cs.unitn.it  |
+----------------------------------------------------------------------+

pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: [HACKERS] New s_lock.h fails on HPUX with gcc
Next
From: Bruce Momjian
Date:
Subject: Change to psql and MB