Re: autovacuum process blocks without reporting a deadlock

yes, u are right.

this are the 3 involved indexes:

hst_timerecording_business_day_idx on hst_timerecording
hst_timerecording_id_employee_idxon hst_timerecording
hst_timerecording_id_timerecording_idxon hst_timerecording

lg t

On Nov 27, 2007 4:07 PM, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> Alvaro Herrera wrote:
> > Thomas Chille wrote:
> > > On Nov 27, 2007 3:14 PM, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> > > hat are the column headings? I find this difficult to read.
> > > >
> > > > Please post the whole of pg_locks. I may be missing something but I
> > > > think we're missing part of the picture here. Autovacuum does not seem
> > > > to be locking on anything.
> > >
> > > Unfortunately i logged the pg_locks-state not well formated. i added
> > > now the heading manually. maybe it is better or i have to wait till
> > > tomorrow morning.
> > >
> > > but what i think i can see, is that the process with pid 9317 is
> > > holding a ShareUpdateExclusiveLock
> >
> > You missed that 9317 is also holding an ExclusiveLock.
>
> Sorry, I misread it. My analysis is bogus :-) I suggest you upgrade
> anyway because there are nasty bugs in the version you are using.
>
> I assume 77862554, 77862552 and 77862553 are indexes on the 75686189
> table?
>
> --
> Alvaro Herrera http://www.amazon.com/gp/registry/CTMLCN8V17R4
>
> "La experiencia nos dice que el hombre peló millones de veces las patatas,
> pero era forzoso admitir la posibilidad de que en un caso entre millones,
> las patatas pelarían al hombre" (Ijon Tichy)
>