> > > Magnus Hagander wrote:
> > >> Per some earlier discussion, here is an attempt at
> implementing a
> > >> "delayed write" of the pgstats file, to decrease the
> write activity
> > >> on that file.
> >
> > This was not ready to be applied, was it? "An attempt"
> doesn't sound
> > to me like Magnus thought it was ready for prime time, and my
> > recollection of the thread is that there were objections.
>
> I don't remember any objection. Magnus, Tom, revert?
I don't recall any specific objections that weren't answered. There was
some (short) talk about having a "backoff" so it won't update the file
too often if the user requests it, but I don't think we ever concluded
if that was necessary or not. If it is, I think we can add that on top
of what's there now, and there's no need to revert.
If there were/are other objections that I missed, a revert might be
needed. Can't really comment since I missed them in that case ;-)
I think the use of "attempt" was more that I expected more comments, and
also somewhat in light of the concurrent discussions about getting rid
of the code to accept delayed destroy messages.
//Magnus
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
