Re: Removing PD_ALL_VISIBLE

On Thu, Jan 17, 2013 at 2:11 PM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
>
> On 17 January 2013 03:02, Jeff Davis <pgsql(at)j-davis(dot)com> wrote:
>
> > Rebased patch attached. No significant changes.
>
> Jeff, can you summarise/collate why we're doing this, what concerns it
> raises and how you've dealt with them? That will help decide whether
> to commit.
>

+1. On another thread "Set visibility map bit after HOT prune", Robert
mentioned that its not such a good idea to remove the PD_ALL_VISIBLE
flag because it helps us to reduce the contention on the VM page,
especially when we need to reset the VM bit. Here is an excerpt from
Robert's comment that thread:

"Sure, but you're zipping rather blithely past the disadvantages of
such an approach. Jeff Davis recently proposed getting rid of
PD_ALL_VISIBLE, and Tom and I both expressed considerable skepticism
about that; this proposal has the same problems. One of the major
benefits of PD_ALL_VISIBLE is that, when it isn't set, inserts,
updates, and deletes to the page can ignore the visibility map. That
means that a server under heavy concurrency is much less likely to
encounter contention on the visibility map blocks. Now, maybe that's
not really a problem, but I sure haven't seen enough evidence to make
me believe it. If it's really true that PD_ALL_VISIBLE needn't fill
this role, then Heikki wasted an awful lot of time implementing it,
and I wasted an awful lot of time keeping it working when I made the
visibility map crash-safe for IOS. That could be true, but I tend to
think it isn't."

May be you've already addressed that concern with the proven
performance numbers, but I'm not sure.

Thanks,
Pavan

--
Pavan Deolasee
http://www.linkedin.com/in/pavandeolasee