I though that following message provided evidence for bug: http://archives.postgresql.org/pgsql-bugs/2007-05/msg00135.php
From: Tom Lane [tgl(at)sss(dot)pgh(dot)pa(dot)us]
Sent: Tuesday, June 07, 2011 7:33 PM
To: Anton Dedov
Subject: Re: [BUGS] ON DELETE CASCADE with multiple paths in PostgreSQL 9.x
Anton Dedov <adedov(at)parallels(dot)com> writes:
> We are using patched version of PostgreSQL 8.1 in order to work-around
> bug described in
> http://archives.postgresql.org/pgsql-bugs/2007-05/msg00130.php. Now, we
> are going to upgrade to 9.0 and it also would be great to get rid of
> patch as it makes keeping database engine current too expensive task.
> I've checked 9.0 and it seems that undesirable behavior is still there.
> Is there any plans to fix it in future releases?
It wasn't a bug then, and it isn't a bug now. There is no guarantee of
the order in which FK constraints will be checked.
regards, tom lane
In response to
pgsql-bugs by date
|Next:||From: Noah Misch||Date: 2011-06-07 19:02:28|
|Subject: Re: BUG #6052: ADD COLUMN - ERROR: tables can have at most 1600columns|
|Previous:||From: Robert Haas||Date: 2011-06-07 18:05:10|
|Subject: Re: BUG #6041: Unlogged table was created bad in slave node|