From: Tom Lane <tgl@sss.pgh.pa.us>
To: Wells Oliver <wells.oliver@gmail.com>
cc: pgsql-admin <pgsql-admin@postgresql.org>
Subject: Re: pg_restore with --use-list and --jobs and depdendencies
In-reply-to: 
 <CAOC+FBV6y-skipCOyrjRxEow1xD9wD-i6PiADTG324_cMpDqaQ@mail.gmail.com>
References: 
 <CAOC+FBV6y-skipCOyrjRxEow1xD9wD-i6PiADTG324_cMpDqaQ@mail.gmail.com>
Comments: In-reply-to Wells Oliver <wells.oliver@gmail.com>
	message dated "Fri, 07 Jan 2022 13:04:14 -0800"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1386275.1641594391.1@sss.pgh.pa.us>
Date: Fri, 07 Jan 2022 17:26:31 -0500
Message-ID: <1386276.1641594391@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1386276.1641594391%40sss.pgh.pa.us>
Precedence: bulk

Wells Oliver <wells.oliver@gmail.com> writes:
> Hi all: I just wanted to clarify something for myself: how does pg_restore
> determine restore order if you're using a --use-list and --jobs parameters.
> Given the concurrency of jobs and (assuming) it goes in order of what's in
> the list, does it... well, does it sort out dependencies properly so data
> is inserted that foreign keys need, etc?

(1) The basic order of the restored items is what you say in the --list
file, but it will start the next item as soon as (a) a worker is free
and (b) all of that item's dependencies are complete

(2) It will honor direct dependencies between restored items, ie
not start B till A is done if B is shown as depending on A

(3) I'm not too sure about indirect dependencies, ie if you have
a situation where C depends on B depends on A and you ask for just
A and C, that may end up with no dependency constraint delaying C.
Which might be fine, or it might not; you'd need to be more specific
about what you intend to skip.

			regards, tom lane