Dietmar Maurer wrote:
> > >
> > > Why cant postgres get the RowExclusiveLock in transaction 3369000?
> >
> > Probably because the ExclusiveLock'ers are waiting in front
> > of RowExclusiveLock. Locks are granted in order.
> >
> > It would help if you didn't mangle the pg_locks output so badly.
>
> Yes, sorry about that.
>
> I was able to reproduce the problem, and the problem is that locks are
> granted in order (wonder why?).
Because doing otherwise would cause starvation for some lockers.
> Anyways, i am trying to avoid locks now, by using my own merge
> function to avoid update/insert race condition.
>
> Or what is the suggested way to avoid the update/insert race condition?.
What update/insert race condition? Maybe you are talking about the
subject of example 37-1 here:
http://www.postgresql.org/docs/current/static/plpgsql-control-structures.html#PLPGSQL-ERROR-TRAPPING
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support