Re: Review: Revise parallel pg_restore's scheduling heuristic

Tom Lane <tgl@sss.pgh.pa.us> wrote: 
> "Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
>> Rebased to correct for pg_indent changes.
> 
> Thanks for doing that.
No problem.  I think I still owe you a few.  :-)
> Attached is a further small improvement that gets rid of the
> find_ready_items() scans.  After re-reading the patch I realized
> that it wasn't *really* avoiding O(N^2) behavior ... but this
> version does.
I'll run a fresh set of benchmarks.
By the way, I muffed the setup of last night's benchmarks, so no new
information there, except that in the process of reviewing the attempt
I discovered I was guilty of making a false assertion yesterday, based
on remembering incorrectly.  The logs show that the six hour dump to
custom format was over the LAN.  :-(  I hope I didn't waste too much
of people's time by saying otherwise.  I'll try to get some numbers on
the same-box dump soon.  (Note to self: never, ever trust your memory;
always confirm.)
-Kevin