To: Simon Riggs <simon@2ndquadrant.com>
cc: Stephen Frost <sfrost@snowman.net>, pgsql-patches@postgresql.org
Subject: Re: pg_dump additional options for performance 
In-reply-to: <1216889809.3894.747.camel@ebony.2ndQuadrant> 
References: <20080720214335.GB18846@tamriel.snowman.net>
	<1216591180.19656.848.camel@ebony.2ndQuadrant>
	<20080721011829.GC18846@tamriel.snowman.net>
	<20080721014942.GC30869@sonic.net>
	<20080721015850.GD18846@tamriel.snowman.net>
	<18415.1216611275@sss.pgh.pa.us>
	<1216621040.19656.867.camel@ebony.2ndQuadrant>
	<20080721115602.GG18846@tamriel.snowman.net>
	<1216831237.3894.676.camel@ebony.2ndQuadrant>
	<1216832987.3894.678.camel@ebony.2ndQuadrant>
	<20080724032047.GI18846@tamriel.snowman.net>
	<1216885031.3894.739.camel@ebony.2ndQuadrant>
	<21687.1216886041@sss.pgh.pa.us>
	<1216889809.3894.747.camel@ebony.2ndQuadrant>
Comments: In-reply-to Simon Riggs <simon@2ndquadrant.com>
	message dated "Thu, 24 Jul 2008 09:56:49 +0100"
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="----- =_aaaaaaaaaa0"
Content-ID: <2697.1217027098.0@sss.pgh.pa.us>
Date: Fri, 25 Jul 2008 19:16:24 -0400
Message-ID: <2818.1217027784@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2697.1217027098.1@sss.pgh.pa.us>

Simon Riggs <simon@2ndquadrant.com> writes:
> [ pg_dump_beforeafter.v6.patch ]

I looked over this patch a bit.  I have a proposal for a slightly
different way of defining the new switches:

* --schema-before-data, --data-only, and --schema-after-data can be
specified in any combination to obtain any subset of the full dump.
If none are specified (which would in itself be a useless combination)
then the default is to dump all three sections, just as if all three
were specified.

* --schema-only is defined as equivalent to specifying both
--schema-before-data and --schema-after-data.

The patch as submitted enforces what seem largely arbitrary restrictions
on combining these switches.  It made some sense before to treat
specifying both --schema-only and --data-only as an error, but it's not
clear to me why you shouldn't be able to write both --schema-before-data
and --schema-after-data, especially when there's a switch right beside
them that appears to be equivalent to that combination.  So let's just
allow all the combinations.

The attached updated patch implements and documents this behavior,
and gets rid of the special linkage between --disable-triggers and
--data-only as previously discussed.

Unfortunately there's still a lot of work to do, and I don't feel like
doing it so I'm bouncing this patch back for further work.

The key problem is that pg_restore is broken: it emits nearly the same
output for --schema-before-data and --schema-after-data, because it
doesn't have any way to distinguish which objects in a full dump file
belong where.  This is because the filtering logic was put in the wrong
place, namely in the ArchiveEntry creation routines in pg_dump.c, when
where it really needs to happen is while scanning the TocEntry list in
RestoreArchive().  (Note: it is perhaps worth keeping the pg_dump.c
filters so as to avoid doing extra server queries for objects that we
aren't going to dump anyway, but the core logic has to be in
RestoreArchive.)

Another issue is that the rules for deciding which objects are "before
data" and which are "after data" are wrong.  In particular ACLs are after
data not before data, which is relatively easy to fix.  Not so easy to fix
is that COMMENTs might be either before or after data depending on what
kind of object they are attached to.

(BTW, what about BLOB COMMENTS?  They definitely can't be "before data".
ISTM you could make a case for them being "after data", if you think that
comments are always schema.  But there is also a case for considering
them as data, because the objects they are attached to are data.  I kind
of like the latter approach because it would create an invariant that
comments appear in the same dump section as the object commented on.
Thoughts?)

Implementing the filtering by examining the type of a TocEntry in 
RestoreArchive is a bit of a PITA, but it's probably possible.  The
main bad thing about that is the need to make an explicit list of every
type of TocEntry that exists now or ever has been emitted by any past
version of pg_dump.  The design concept was that the type tags are
mainly documentation, and while we've had to bend that in places (mostly
for backward-compatibility reasons) this would be the first place we'd
have to throw it overboard completely.

And there's yet another issue here, which is that it's not entirely clear
that the type of an object uniquely determines whether it's before or
after data.  This might be an emergent property of the object sorting
rules, but there is certainly not anything positively guaranteeing that
the dependency-driven topological sort will produce such a result, and
especially not that that will always be true in the future.  So the
approach seems a bit fragile.

We could perhaps get rid of that problem, as well as the need to implement
object-type-determination logic, if we were to make RestoreArchive define
the groupings according to position in the TocEntry list: everything
before the first TABLE DATA or BLOB (and BLOB COMMENT?) entry is "before"
data, everything after the last one is "after" data, everything in between
is data.  Then we only need to identify object types that are considered
"data", which we already have a rule for (whether hadDumper is true).
This is pretty attractive until you stop to consider the possibility
that there aren't any data entries in an archive (ie, it was made with
--schema-only): then there's no way to identify the boundary points.

We could solve that problem by inserting a "dummy data" TOC entry where
the data would have appeared, but this will only work in new archive
files.  With an implementation like this, pg_restore with
--schema-before-data or --schema-after-data won't behave very nicely on a
pre-8.4 --schema-only archive file.  (Presumably it would act as though
all the objects were "before" data.)  Is that a small enough corner case
to live with in order to gain implementation simplicity and robustness?

BTW, another incomplete item is that pg_dumpall should probably be taught
to accept and pass down --schema-before-data and --schema-after-data
switches.

			regards, tom lane


------- =_aaaaaaaaaa0
Content-Type: application/octet-stream
Content-ID: <2697.1217027098.2@sss.pgh.pa.us>
Content-Description: pg_dump_beforeafter.v7.patch.gz
Content-Transfer-Encoding: base64
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------- =_aaaaaaaaaa0--