Ron <ronljohnsonjr(at)gmail(dot)com> writes:
> On 08/26/2018 01:42 PM, Tom Lane wrote:
>> Perhaps I don't understand *your* question. What concrete problem are you
>> having?
> I want to track the progress of pg_dump so as to estimate completion time.
Well, if you don't use --jobs then you should get more or less the same
behavior as in 8.4. If you do use that, then it's hardly surprising that
things are more complicated.
It looks like there's a hack in there to order tables by decreasing size
during a parallel dump, on the theory that we'll get more parallelism
that way. (Not sure I believe that, and the part of it that also orders
index dumps that way is certainly brain-dead.)
regards, tom lane