From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Philip de Nier <philip(dot)denier(at)rd(dot)bbc(dot)co(dot)uk> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: using deferred initially deferred to solve foreign key checking issues |
Date: | 2006-08-22 21:21:12 |
Message-ID: | 11844.1156281672@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Philip de Nier <philip(dot)denier(at)rd(dot)bbc(dot)co(dot)uk> writes:
> The solutions I could find in the mailing lists were either to upgrade to
> version 8.1 which uses "SELECT ... FOR SHARE" (I'm currently using 8.0),
> stop using foreign keys or add "DEFERRABLE INITIALLY DEFERRED" to the
> constraints. For now I'd prefer to use the last option.
> Does using "DEFERRABLE INITIALLY DEFERRED" completely solve the problem of
> transactions waiting for each other to release locks resulting from "SELECT
> ... FOR UPDATE"?
It doesn't eliminate the problem, but it does narrow the window in which
the lock is held by quite a bit, by postponing the FK checks until just
before transaction commit.
> How are the deferred foreign key constraints checked - are they checked one
> at a time?
Yeah.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2006-08-22 21:24:10 | Re: error msg when pg_restore |
Previous Message | Tom Lane | 2006-08-22 21:09:51 | Re: share library version problems |