On Wed, Dec 19, 2012 at 9:51 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> If we start generating a lot of useless WAL activity and I/O as
> a result of thrashing the all-visible bit, it won't be so tolerable
> anymore.
What if we wrap that into the WAL generated by HOT prune itself ?
Would that address your concerns for extra WAL logging ? I also
suggested doing it conditionally to avoid contention on the VM buffer.
(I actually wonder why we WAL-log set operation at all except for HS
to be able to do IOS, but thats a topic for separate thread may be)
Also, if extra WAL-logging is really worrisome, may be we should again
seriously reconsider my idea of *not* clearing the bit at HOT update.
My apologies for repeating myself. But that seems very important in a
steady system when almost every update is a HOT update. So you don't
clear the bit at HOT update and so don't need to set it back either,
thus saving two WAL activity. You definitely don't need any vacuum in
this case if pruning keeps reclaiming dead space at appropriate time
and make it available for the next update. More so, IOS still works
great. Why is this so bad ? I haven't forgotten your complaints about
changed meaning of the bit, but I tried to explain that we can read it
in a slightly different way and still show it as an invariant.
>
> I think my core point still stands: the way that HOT pruning is done now
> is an artifact of having wanted to shoehorn it into the system with
> minimum changes. Which was reasonable at the time given the
> experimental status of the feature, but now it's time to reconsider.
>
ISTM that you already have concret ideas about what are those places
where HOT prune would be more effective. My worry is changing anything
there is going to be a lot trickier and will require heavy testing.
Our initial work has served us well so far. Of course, I've no problem
changing that if its going to benefit users.
Thanks,
Pavan
--
Pavan Deolasee
http://www.linkedin.com/in/pavandeolasee