From: Tom Lane <tgl@sss.pgh.pa.us>
To: Stephen Frost <sfrost@snowman.net>
cc: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>,
        pgsql-hackers <pgsql-hackers@postgresql.org>
Subject: Re: pg_dump test instability
In-reply-to: <20180827144138.GO3326@tamriel.snowman.net>
References: <5137fe12-d0a2-4971-61b6-eb4e7e8875f8@2ndquadrant.com>
 <32399.1535379095@sss.pgh.pa.us> <20180827144138.GO3326@tamriel.snowman.net>
Comments: In-reply-to Stephen Frost <sfrost@snowman.net>
	message dated "Mon, 27 Aug 2018 10:41:38 -0400"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1664.1535381158.1@sss.pgh.pa.us>
Date: Mon, 27 Aug 2018 10:45:58 -0400
Message-ID: <1665.1535381158@sss.pgh.pa.us>
Precedence: bulk

Stephen Frost <sfrost@snowman.net> writes:
> * Tom Lane (tgl@sss.pgh.pa.us) wrote:
>> However, at least for the directory-format case (which I think is the
>> only one supported for parallel restore), we could make it compare the
>> file sizes of the TABLE DATA items.  That'd work pretty well as a proxy
>> for both the amount of effort needed for table restore, and the amount
>> of effort needed to build indexes on the tables afterwards.

> Parallel restore also works w/ custom-format dumps.

Really.  Well then the existing code is even more broken, because it
only does this sorting for directory output:

	/* If we do a parallel dump, we want the largest tables to go first */
	if (archiveFormat == archDirectory && numWorkers > 1)
		sortDataAndIndexObjectsBySize(dobjs, numObjs);

so that parallel restore is completely left in the lurch with a
custom-format dump.

But I imagine we can get some measure of table data size out of a custom
dump too.

			regards, tom lane