Commit c0d5be5d6 caused pg_dump to create a separate BLOB metadata TOC
entry for each large object (blob), but it did not touch the ancient
decision to put all the blobs' data into a single "BLOBS" TOC entry.
This is bad for a few reasons: for databases with millions of blobs,
the TOC becomes unreasonably large, causing performance issues;
selective restore of just some blobs is quite impossible; and we
cannot parallelize either dump or restore of the blob data, since our
architecture for that relies on farming out whole TOC entries to
worker processes.
To improve matters, let's group multiple blobs into each blob metadata
TOC entry, and then make corresponding per-group blob data TOC entries.
Selective restore using pg_restore's -l/-L switches is then possible,
though only at the group level. (Perhaps we should provide a switch
to allow forcing one-blob-per-group for users who need precise
selective restore and don't have huge numbers of blobs. This patch
doesn't do that, instead just hard-wiring the maximum number of blobs
per entry at 1000.)
The blobs in a group must all have the same owner, since the TOC entry
format only allows one owner to be named. In this implementation
we also require them to all share the same ACL (grants); the archive
format wouldn't require that, but pg_dump's representation of
DumpableObjects does. It seems unlikely that either restriction
will be problematic for databases with huge numbers of blobs.
The metadata TOC entries now have a "desc" string of "BLOB METADATA",
and their "defn" string is just a newline-separated list of blob OIDs.
The restore code has to generate creation commands, ALTER OWNER
commands, and drop commands (for --clean mode) from that. We would
need special-case code for ALTER OWNER and drop in any case, so the
alternative of keeping the "defn" as directly executable SQL code
for creation wouldn't buy much, and it seems like it'd bloat the
archive to little purpose.
Since we require the blobs of a metadata group to share the same ACL,
we can furthermore store only one copy of that ACL, and then make
pg_restore regenerate the appropriate commands for each blob. This
saves space in the dump file not only by removing duplicative SQL
command strings, but by not needing a separate TOC entry for each
blob's ACL. In turn, that reduces client-side memory requirements for
handling many blobs.
ACL TOC entries that need this special processing are labeled as
"ACL"/"LARGE OBJECTS nnn..nnn". If we have a blob with a unique ACL,
continue to label it as "ACL"/"LARGE OBJECT nnn". We don't actually
have to make such a distinction, but it saves a few cycles during
restore for the easy case, and it seems like a good idea to not change
the TOC contents unnecessarily.
The data TOC entries ("BLOBS") are exactly the same as before,
except that now there can be more than one, so we'd better give them
identifying tag strings.
Also, commit c0d5be5d6 put the new BLOB metadata TOC entries into
SECTION_PRE_DATA, which perhaps is defensible in some ways, but
it's a rather odd choice considering that we go out of our way to
treat blobs as data. Moreover, because parallel restore handles
the PRE_DATA section serially, this means we'd only get part of the
parallelism speedup we could hope for. Move these entries into
SECTION_DATA, letting us parallelize the lo_create calls not just the
data loading when there are many blobs. Add dependencies to ensure
that we won't try to load data for a blob we've not yet created.
As this stands, we still generate a separate TOC entry for any comment
or security label attached to a blob. I feel comfortable in believing
that comments and security labels on blobs are rare, so this patch
should be enough to get most of the useful TOC compression for blobs.
We have to bump the archive file format version number, since existing
versions of pg_restore wouldn't know they need to do something special
for BLOB METADATA, plus they aren't going to work correctly with
multiple BLOBS entries or multiple-large-object ACL entries.
The directory and tar-file format handlers need some work
for multiple BLOBS entries: they used to hard-wire the file name
as "blobs.toc", which is replaced here with "blobs_<dumpid>.toc".
The 002_pg_dump.pl test script also knows about that and requires
minor updates. (I had to drop the test for manually-compressed
blobs.toc files with LZ4, because lz4's obtuse command line
design requires explicit specification of the output file name
which seems impractical here. I don't think we're losing any
useful test coverage thereby; that test stanza seems completely
duplicative with the gzip and zstd cases anyway.)
In passing, centralize management of the lo_buf used to hold data
while restoring blobs. The code previously had each format handler
create lo_buf, which seems rather pointless given that the format
handlers all make it the same way. Moreover, the format handlers
never use lo_buf directly, making this setup a failure from a
separation-of-concerns standpoint. Let's move the responsibility into
pg_backup_archiver.c, which is the only module concerned with lo_buf.
The reason to do this in this patch is that it allows a centralized
fix for the now-false assumption that we never restore blobs in
parallel. Also, get rid of dead code in DropLOIfExists: it's been a
long time since we had any need to be able to restore to a pre-9.0
server.
Discussion: https://postgr.es/m/a9f9376f1c3343a6bb319dce294e20ac@EX13D05UWC001.ant.amazon.com
This feature will allow us to replicate the changes on subscriber nodes
after the upgrade.
Previously, only the subscription metadata information was preserved.
Without the list of relations and their state, it's not possible to
re-enable the subscriptions without missing some records as the list of
relations can only be refreshed after enabling the subscription (and
therefore starting the apply worker). Even if we added a way to refresh
the subscription while enabling a publication, we still wouldn't know
which relations are new on the publication side, and therefore should be
fully synced, and which shouldn't.
To preserve the subscription relations, this patch teaches pg_dump to
restore the content of pg_subscription_rel from the old cluster by using
binary_upgrade_add_sub_rel_state SQL function. This is supported only
in binary upgrade mode.
The subscription's replication origin is needed to ensure that we don't
replicate anything twice.
To preserve the replication origins, this patch teaches pg_dump to update
the replication origin along with creating a subscription by using
binary_upgrade_replorigin_advance SQL function to restore the
underlying replication origin remote LSN. This is supported only in
binary upgrade mode.
pg_upgrade will check that all the subscription relations are in 'i'
(init) or in 'r' (ready) state and will error out if that's not the case,
logging the reason for the failure. This helps to avoid the risk of any
dangling slot or origin after the upgrade.
Author: Vignesh C, Julien Rouhaud, Shlok Kyal
Reviewed-by: Peter Smith, Masahiko Sawada, Michael Paquier, Amit Kapila, Hayato Kuroda
Discussion: https://postgr.es/m/20230217075433.u5mjly4d5cr4hcfe@jrouhaud
We now create contype='n' pg_constraint rows for not-null constraints.
We propagate these constraints to other tables during operations such as
adding inheritance relationships, creating and attaching partitions and
creating tables LIKE other tables. We also spawn not-null constraints
for inheritance child tables when their parents have primary keys.
These related constraints mostly follow the well-known rules of
conislocal and coninhcount that we have for CHECK constraints, with some
adaptations: for example, as opposed to CHECK constraints, we don't
match not-null ones by name when descending a hierarchy to alter it,
instead matching by column name that they apply to. This means we don't
require the constraint names to be identical across a hierarchy.
For now, we omit them for system catalogs. Maybe this is worth
reconsidering. We don't support NOT VALID nor DEFERRABLE clauses
either; these can be added as separate features later (this patch is
already large and complicated enough.)
psql shows these constraints in \d+.
pg_dump requires some ad-hoc hacks, particularly when dumping a primary
key. We now create one "throwaway" not-null constraint for each column
in the PK together with the CREATE TABLE command, and once the PK is
created, all those throwaway constraints are removed. This avoids
having to check each tuple for nullness when the dump restores the
primary key creation.
pg_upgrading from an older release requires a somewhat brittle procedure
to create a constraint state that matches what would be created if the
database were being created fresh in Postgres 17. I have tested all the
scenarios I could think of, and it works correctly as far as I can tell,
but I could have neglected weird cases.
This patch has been very long in the making. The first patch was
written by Bernd Helmle in 2010 to add a new pg_constraint.contype value
('n'), which I (Álvaro) then hijacked in 2011 and 2012, until that one
was killed by the realization that we ought to use contype='c' instead:
manufactured CHECK constraints. However, later SQL standard
development, as well as nonobvious emergent properties of that design
(mostly, failure to distinguish them from "normal" CHECK constraints as
well as the performance implication of having to test the CHECK
expression) led us to reconsider this choice, so now the current
implementation uses contype='n' again. During Postgres 16 this had
already been introduced by commit e056c557ae, but there were some
problems mainly with the pg_upgrade procedure that couldn't be fixed in
reasonable time, so it was reverted.
In 2016 Vitaly Burovoy also worked on this feature[1] but found no
consensus for his proposed approach, which was claimed to be closer to
the letter of the standard, requiring an additional pg_attribute column
to track the OID of the not-null constraint for that column.
[1] https://postgr.es/m/CAKOSWNkN6HSyatuys8xZxzRCR-KL1OkHS5-b9qd9bf1Rad3PLA@mail.gmail.com
Author: Álvaro Herrera <alvherre@alvh.no-ip.org>
Author: Bernd Helmle <mailings@oopsware.de>
Reviewed-by: Justin Pryzby <pryzby@telsasoft.com>
Reviewed-by: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Reviewed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
This reverts commit e056c557ae and minor later fixes thereof.
There's a few problems in this new feature -- most notably regarding
pg_upgrade behavior, but others as well. This new feature is not in any
way critical on its own, so instead of scrambling to fix it we revert it
and try again in early 17 with these issues in mind.
Discussion: https://postgr.es/m/3801207.1681057430@sss.pgh.pa.us
We now create pg_constaint rows for NOT NULL constraints with
contype='n'.
We propagate these constraints during operations such as adding
inheritance relationships, creating and attaching partitions, creating
tables LIKE other tables. We mostly follow the well-known rules of
conislocal and coninhcount that we have for CHECK constraints, with some
adaptations; for example, as opposed to CHECK constraints, we don't
match NOT NULL ones by name when descending a hierarchy to alter it;
instead we match by column number. This means we don't require the
constraint names to be identical across a hierarchy.
For now, we omit them from system catalogs. Maybe this is worth
reconsidering. We don't support NOT VALID nor DEFERRABLE clauses
either; these can be added as separate features later (this patch is
already large and complicated enough.)
This has been very long in the making. The first patch was written by
Bernd Helmle in 2010 to add a new pg_constraint.contype value ('n'),
which I (Álvaro) then hijacked in 2011 and 2012, until that one was
killed by the realization that we ought to use contype='c' instead:
manufactured CHECK constraints. However, later SQL standard
development, as well as nonobvious emergent properties of that design
(mostly, failure to distinguish them from "normal" CHECK constraints as
well as the performance implication of having to test the CHECK
expression) led us to reconsider this choice, so now the current
implementation uses contype='n' again.
In 2016 Vitaly Burovoy also worked on this feature[1] but found no
consensus for his proposed approach, which was claimed to be closer to
the letter of the standard, requiring additional pg_attribute columns to
track the OID of the NOT NULL constraint for that column.
[1] https://postgr.es/m/CAKOSWNkN6HSyatuys8xZxzRCR-KL1OkHS5-b9qd9bf1Rad3PLA@mail.gmail.com
Author: Álvaro Herrera <alvherre@alvh.no-ip.org>
Author: Bernd Helmle <mailings@oopsware.de>
Reviewed-by: Justin Pryzby <pryzby@telsasoft.com>
Reviewed-by: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Discussion: https://postgr.es/m/CACA0E642A0267EDA387AF2B%40%5B172.26.14.62%5D
Discussion: https://postgr.es/m/AANLkTinLXMOEMz+0J29tf1POokKi4XDkWJ6-DDR9BKgU@mail.gmail.com
Discussion: https://postgr.es/m/20110707213401.GA27098@alvh.no-ip.org
Discussion: https://postgr.es/m/1343682669-sup-2532@alvh.no-ip.org
Discussion: https://postgr.es/m/CAKOSWNkN6HSyatuys8xZxzRCR-KL1OkHS5-b9qd9bf1Rad3PLA@mail.gmail.com
Discussion: https://postgr.es/m/20220817181249.q7qvj3okywctra3c@alvherre.pgsql
Instead of trying to optimize this by skipping creation of the
links for tables we don't plan to dump, just create them all in
bulk with a single scan over the pg_inherits data. The previous
approach was more or less O(N^2) in the number of pg_inherits
entries, not to mention being way too complicated.
Also, don't create useless TableAttachInfo objects.
It's silly to create a TableAttachInfo object that we're not
going to dump, when we know perfectly well at creation time
that it won't be dumped.
Patch by me; thanks to Julien Rouhaud for review.
Discussion: https://postgr.es/m/1376149.1675268279@sss.pgh.pa.us
Hash partitioning on an enum is problematic because the hash codes are
derived from the OIDs assigned to the enum values, which will almost
certainly be different after a dump-and-reload than they were before.
This means that some rows probably end up in different partitions than
before, causing restore to fail because of partition constraint
violations. (pg_upgrade dodges this problem by using hacks to force
the enum values to keep the same OIDs, but that's not possible nor
desirable for pg_dump.)
Users can work around that by specifying --load-via-partition-root,
but since that's a dump-time not restore-time decision, one might
find out the need for it far too late. Instead, teach pg_dump to
apply that option automatically when dealing with a partitioned
table that has hash-on-enum partitioning.
Also deal with a pre-existing issue for --load-via-partition-root
mode: in a parallel restore, we try to TRUNCATE target tables just
before loading them, in order to enable some backend optimizations.
This is bad when using --load-via-partition-root because (a) we're
likely to suffer deadlocks from restore jobs trying to restore rows
into other partitions than they came from, and (b) if we miss getting
a deadlock we might still lose data due to a TRUNCATE removing rows
from some already-completed restore job.
The fix for this is conceptually simple: just don't TRUNCATE if we're
dealing with a --load-via-partition-root case. The tricky bit is for
pg_restore to identify those cases. In dumps using COPY commands we
can inspect each COPY command to see if it targets the nominal target
table or some ancestor. However, in dumps using INSERT commands it's
pretty impractical to examine the INSERTs in advance. To provide a
solution for that going forward, modify pg_dump to mark TABLE DATA
items that are using --load-via-partition-root with a comment.
(This change also responds to a complaint from Robert Haas that
the dump output for --load-via-partition-root is pretty confusing.)
pg_restore checks for the special comment as well as checking the
COPY command if present. This will fail to identify the combination
of --load-via-partition-root and --inserts in pre-existing dump files,
but that should be a pretty rare case in the field. If it does
happen you will probably get a deadlock failure that you can work
around by not using parallel restore, which is the same as before
this bug fix.
Having done this, there seems no remaining reason for the alarmism
in the pg_dump man page about combining --load-via-partition-root
with parallel restore, so remove that warning.
Patch by me; thanks to Julien Rouhaud for review. Back-patch to
v11 where hash partitioning was introduced.
Discussion: https://postgr.es/m/1376149.1675268279@sss.pgh.pa.us
As it stands, flagInhAttrs() can make changes in table properties that
change decisions made at other tables during other iterations of its
loop. This is a pretty bad idea, since we visit the tables in OID
order which is not necessarily related to inheritance relationships.
So far as I can tell, the consequences are just cosmetic: we might
dump DEFAULT or GENERATED expressions that we don't really need to
because they match properties of the parent. Nonetheless, it's buggy,
and somebody might someday add functionality here that fails less
benignly when the traversal order varies.
One issue is that when we decide we needn't dump a particular
GENERATED expression, we physically unlink the struct for it,
so that it will now look like the table has no such expression,
causing the wrong choice to be made at any child visited later.
We can improve that by instead clearing the dobj.dump flag,
and taking care to check that flag when it comes time to dump
the expression or not.
The other problem is that if we decide we need to fake up a DEFAULT
NULL clause to override a default that would otherwise get inherited,
we modify the data structure in the reverse fashion, creating an
attrdefs entry where there hadn't been one. It's harder to avoid
doing that, but since the backend won't report a plain "DEFAULT NULL"
property we can modify the code to recognize ones we just added.
Add some commentary to perhaps forestall future mistakes of the
same ilk.
Since the effects of this seem only cosmetic, no back-patch.
Discussion: https://postgr.es/m/1506298.1676323579@sss.pgh.pa.us
In both partitioning and traditional inheritance, require child
columns to be GENERATED if and only if their parent(s) are.
Formerly we allowed the case of an inherited column being
GENERATED when its parent isn't, but that results in inconsistent
behavior: the column can be directly updated through an UPDATE
on the parent table, leading to it containing a user-supplied
value that might not match the generation expression. This also
fixes an oversight that we enforced partition-key-columns-can't-
be-GENERATED against parent tables, but not against child tables
that were dynamically attached to them.
Also, remove the restriction that the child's generation expression
be equivalent to the parent's. In the wake of commit 3f7836ff6,
there doesn't seem to be any reason that we need that restriction,
since generation expressions are always computed per-table anyway.
By removing this, we can also allow a child to merge multiple
inheritance parents with inconsistent generation expressions, by
overriding them with its own expression, much as we've long allowed
for DEFAULT expressions.
Since we're rejecting a case that we used to accept, this doesn't
seem like a back-patchable change. Given the lack of field
complaints about the inconsistent behavior, it's likely that no
one is doing this anyway, but we won't change it in minor releases.
Amit Langote and Tom Lane
Discussion: https://postgr.es/m/2793383.1672944799@sss.pgh.pa.us
Make sure that function declarations use names that exactly match the
corresponding names from function definitions in pg_dump/pg_dumpall
related code.
Affected code happens to be inconsistent in how it applies conventions
around how Archive and Archive Handle variables are named. Significant
code churn is required to fully fix those inconsistencies, so take the
least invasive approach possible: treat function definition names as
authoritative, and mechanically adjust corresponding names from function
definitions to match.
Like other recent commits that cleaned up function parameter names, this
commit was written with help from clang-tidy.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/CAH2-Wzmma+vzcO6gr5NYDZ+sx0G14aU-UrzFutT2FoRaisVCUQ@mail.gmail.com
Get rid of the separate "FATAL" log level, as it was applied
so inconsistently as to be meaningless. This mostly involves
s/pg_log_fatal/pg_log_error/g.
Create a macro pg_fatal() to handle the common use-case of
pg_log_error() immediately followed by exit(1). Various
modules had already invented either this or equivalent macros;
standardize on pg_fatal() and apply it where possible.
Invent the ability to add "detail" and "hint" messages to a
frontend message, much as we have long had in the backend.
Except where rewording was needed to convert existing coding
to detail/hint style, I have (mostly) resisted the temptation
to change existing message wording.
Patch by me. Design and patch reviewed at various stages by
Robert Haas, Kyotaro Horiguchi, Peter Eisentraut and
Daniel Gustafsson.
Discussion: https://postgr.es/m/1363732.1636496441@sss.pgh.pa.us
Split the DumpableObject.dump bitmask field into separate bitmasks
tracking which components are requested to be dumped (in the
existing "dump" field) and which components exist for the particular
object (in the new "components" field). This gets rid of some
klugy and easily-broken logic that involved setting bits and later
clearing them. More importantly, it restores the originally intended
behavior that pg_dump's secondary data-gathering queries should not
be executed for objects we have no interest in dumping. That
optimization got broken when the dump flag was turned into a bitmask,
because irrelevant bits tended to remain set in many cases. Since
the "components" field starts from a minimal set of bits and is
added onto as needed, ANDing it with "dump" provides a reliable
indicator of what we actually have to dump, without having to
complicate the logic that manages the request bits. This makes
a significant difference in the number of queries needed when,
for example, there are many functions in extensions.
Discussion: https://postgr.es/m/2273648.1634764485@sss.pgh.pa.us
Discussion: https://postgr.es/m/7d7eb6128f40401d81b3b7a898b6b4de@W2012-02.nidsa.loc
A new option "FOR ALL TABLES IN SCHEMA" in Create/Alter Publication allows
one or more schemas to be specified, whose tables are selected by the
publisher for sending the data to the subscriber.
The new syntax allows specifying both the tables and schemas. For example:
CREATE PUBLICATION pub1 FOR TABLE t1,t2,t3, ALL TABLES IN SCHEMA s1,s2;
OR
ALTER PUBLICATION pub1 ADD TABLE t1,t2,t3, ALL TABLES IN SCHEMA s1,s2;
A new system table "pg_publication_namespace" has been added, to maintain
the schemas that the user wants to publish through the publication.
Modified the output plugin (pgoutput) to publish the changes if the
relation is part of schema publication.
Updates pg_dump to identify and dump schema publications. Updates the \d
family of commands to display schema publications and \dRp+ variant will
now display associated schemas if any.
Author: Vignesh C, Hou Zhijie, Amit Kapila
Syntax-Suggested-by: Tom Lane, Alvaro Herrera
Reviewed-by: Greg Nancarrow, Masahiko Sawada, Hou Zhijie, Amit Kapila, Haiying Tang, Ajin Cherian, Rahila Syed, Bharath Rupireddy, Mark Dilger
Tested-by: Haiying Tang
Discussion: https://www.postgresql.org/message-id/CALDaNm0OANxuJ6RXqwZsM1MSY4s19nuH3734j4a72etDwvBETQ@mail.gmail.com
I found these by running pg_dump under "valgrind --leak-check=full".
The changes in flagInhIndexes() and getIndexes() replace allocation of
an array of which we use only some elements by individual allocations
of just the actually-needed objects. The previous coding wasted some
memory, but more importantly it confused valgrind's leak tracking.
collectComments() and collectSecLabels() remain major blots on
the valgrind report, because they don't PQclear their query
results, in order to avoid a lot of strdup's. That's a dubious
tradeoff, but I'll leave it alone here; an upcoming patch will
modify those functions enough to justify changing the tradeoff.
Create a hash table that indexes dumpable objects by CatalogId
(that is, catalog OID + object OID). Use this to replace the
former catalogIdMap array, as well as various other single-
catalog index arrays, and also the extension membership map.
In principle this should be faster for databases with many objects,
since lookups are now O(1) not O(log N). However, it seems that these
lookups are pretty much negligible in context, so that no overall
performance change can be measured. But having only one lookup
data structure to maintain makes the code simpler and more flexible,
so let's do it anyway.
Discussion: https://postgr.es/m/2595220.1634855245@sss.pgh.pa.us
Also "make reformat-dat-files".
The only change worthy of note is that pgindent messed up the formatting
of launcher.c's struct LogicalRepWorkerId, which led me to notice that
that struct wasn't used at all anymore, so I just took it out.
Generation expressions of generated columns are always inherited, so
there is no need to set them separately in child tables, and there is
no syntax to do so either. The code previously used the code paths
for the handling of default values, for which different rules apply;
in particular it might want to set a default value explicitly for an
inherited column. This resulted in unrestorable dumps. For generated
columns, just skip them in inherited tables.
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://www.postgresql.org/message-id/flat/15830.1575468847%40sss.pgh.pa.us
This is the same fix as commit 9eabfe300 applied to INDEX ATTACH
entries, but for table-to-publication attachments. As in that
case, even though the backend doesn't record "ownership" of the
attachment, we still ought to label it in the dump archive with
the role name that should run the ALTER PUBLICATION command.
The existing behavior causes the ALTER to be done by the original
role that started the restore; that will usually work fine, but
there may be corner cases where it fails.
The bulk of the patch is concerned with changing struct
PublicationRelInfo to include a pointer to the associated
PublicationInfo object, so that we can get the owner's name
out of that when the time comes. While at it, I rewrote
getPublicationTables() to do just one query of pg_publication_rel,
not one per table.
Back-patch to v10 where this code was introduced.
Discussion: https://postgr.es/m/1165710.1610473242@sss.pgh.pa.us
Previously, we emitted the ATTACH PARTITION command as part of
the child table's ArchiveEntry. This was a poor choice since it
complicates restoring the partition as a standalone table; you have
to ignore the error from the ATTACH, which isn't even an option when
restoring direct-to-database with pg_restore. (pg_restore will issue
the whole ArchiveEntry as one PQexec, so that any error rolls back
the table creation as well.) Hence, separate it out as its own
ArchiveEntry, as indeed we already did for index ATTACH PARTITION
commands.
Justin Pryzby
Discussion: https://postgr.es/m/20201023052940.GE9241@telsasoft.com
If there are no objects of a certain type, there is no need to do an
allocation for a set of DumpableObject items. The previous coding did
an allocation of 1 byte instead as per the fallback of pg_malloc() in
the event of an allocation size of zero. This assigns NULL instead for
a set of dumpable objects.
A similar rule already applied to findObjectByOid(), so this makes the
code more defensive as we would just fail with a pointer dereference
instead of attempting to use some incorrect data if a non-existing,
positive, OID is given by a caller of this function.
Author: Daniel Gustafsson
Reviewed-by: Julien Rouhaud, Ranier Vilela
Discussion: https://postgr.es/m/26C43E58-BDD0-4F1A-97CC-4A07B52E32C5@yesql.se
Parallel pg_restore has always supposed that ACL items for different
objects are independent and can be restored in parallel without
conflicts. However, there is one case where this fails: because
REVOKE on a table is defined to also revoke the privilege(s) at
column level, we can't restore per-column ACLs till after we restore
any table-level privileges on their table. Failure to honor this
restriction can lead to "tuple concurrently updated" errors during
parallel restore, or even to the per-column ACLs silently disappearing
because the table-level REVOKE is executed afterwards.
To fix, add a dependency from each column-level ACL item to its table's
ACL item, if there is one. Note that this doesn't fix the hazard
for pre-existing archive files, only for ones made with a corrected
pg_dump. Given that the bug's been there quite awhile without
field reports, I think this is acceptable.
This requires changing the API of pg_dump's dumpACL() function.
To keep its argument list from getting even longer, I removed the
"CatalogId objCatId" argument, which has been unused for ages.
Per report from Justin Pryzby. Back-patch to all supported branches.
Discussion: https://postgr.es/m/20200706050129.GW4107@telsasoft.com
pg_dump is oblivious to this kind of dependency, so they're lost on
dump/restores (and pg_upgrade). Have pg_dump emit ALTER lines so that
they're preserved. Add some pg_dump tests for the whole thing, also.
Reviewed-by: Tom Lane (offlist)
Reviewed-by: Ibrar Ahmed
Reviewed-by: Ahsan Hadi (who also reviewed commit 899a04f5ed)
Discussion: https://postgr.es/m/20200217225333.GA30974@alvherre.pgsql
When an FK constraint is created, it needs the index on the referenced
table to exist and be valid. When doing parallel pg_restore and the
referenced table was partitioned, this condition can sometimes not be
met, because pg_dump didn't emit sufficient object dependencies to
ensure so; this means that parallel pg_restore would fail in certain
conditions. Fix by having pg_dump make the FK constraint object
dependent on the partition attachment objects for the constraint's
referenced index.
This has been broken since f56f8f8da6, so backpatch to Postgres 12.
Discussion: https://postgr.es/m/20191005224333.GA9738@alvherre.pgsql
This is still using the 2.0 version of pg_bsd_indent.
I thought it would be good to commit this separately,
so as to document the differences between 2.0 and 2.1 behavior.
Discussion: https://postgr.es/m/16296.1558103386@sss.pgh.pa.us
The original placement of this module in src/fe_utils/ is ill-considered,
because several src/common/ modules have dependencies on it, meaning that
libpgcommon and libpgfeutils now have mutual dependencies. That makes it
pointless to have distinct libraries at all. The intended design is that
libpgcommon is lower-level than libpgfeutils, so only dependencies from
the latter to the former are acceptable.
We already have the precedent that fe_memutils and a couple of other
modules in src/common/ are frontend-only, so it's not stretching anything
out of whack to treat logging.c as a frontend-only module in src/common/.
To the extent that such modules help provide a common frontend/backend
environment for the rest of common/ to use, it's a reasonable design.
(logging.c does not yet provide an ereport() emulation, but one can
dream.)
Hence, move these files over, and revert basically all of the build-system
changes made by commit cc8d41511. There are no places that need to grow
new dependencies on libpgcommon, further reinforcing the idea that this
is the right solution.
Discussion: https://postgr.es/m/a912ffff-f6e4-778a-c86a-cf5c47a12933@2ndquadrant.com
This unifies the various ad hoc logging (message printing, error
printing) systems used throughout the command-line programs.
Features:
- Program name is automatically prefixed.
- Message string does not end with newline. This removes a common
source of inconsistencies and omissions.
- Additionally, a final newline is automatically stripped, simplifying
use of PQerrorMessage() etc., another common source of mistakes.
- I converted error message strings to use %m where possible.
- As a result of the above several points, more translatable message
strings can be shared between different components and between
frontends and backend, without gratuitous punctuation or whitespace
differences.
- There is support for setting a "log level". This is not meant to be
user-facing, but can be used internally to implement debug or
verbose modes.
- Lazy argument evaluation, so no significant overhead if logging at
some level is disabled.
- Some color in the messages, similar to gcc and clang. Set
PG_COLOR=auto to try it out. Some colors are predefined, but can be
customized by setting PG_COLORS.
- Common files (common/, fe_utils/, etc.) can handle logging much more
simply by just using one API without worrying too much about the
context of the calling program, requiring callbacks, or having to
pass "progname" around everywhere.
- Some programs called setvbuf() to make sure that stderr is
unbuffered, even on Windows. But not all programs did that. This
is now done centrally.
Soft goals:
- Reduces vertical space use and visual complexity of error reporting
in the source code.
- Encourages more deliberate classification of messages. For example,
in some cases it wasn't clear without analyzing the surrounding code
whether a message was meant as an error or just an info.
- Concepts and terms are vaguely aligned with popular logging
frameworks such as log4j and Python logging.
This is all just about printing stuff out. Nothing affects program
flow (e.g., fatal exits). The uses are just too varied to do that.
Some existing code had wrappers that do some kind of print-and-exit,
and I adapted those.
I tried to keep the output mostly the same, but there is a lot of
historical baggage to unwind and special cases to consider, and I
might not always have succeeded. One significant change is that
pg_rewind used to write all error messages to stdout. That is now
changed to stderr.
Reviewed-by: Donald Dong <xdong@csumb.edu>
Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru>
Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
Ensure the TOC entry is marked with the correct schema, so that its
name is as unique as the index's is.
Fix the dependencies: we want dependencies from this TOC entry to the
two indexes it depends on, and we don't care (at least not for this
purpose) what order the indexes are created in. Also, add dependencies
on the indexes' underlying tables. Those might seem pointless given
the index dependencies, but they are helpful to cue parallel restore
to avoid running the ATTACH PARTITION in parallel with other DDL on
the same tables.
Discussion: https://postgr.es/m/10817.1535494963@sss.pgh.pa.us
Everything of use to frontend code should now appear in the _d.h files,
and making this change frees us from needing to worry about whether the
catalog header files proper are frontend-safe.
Remove src/interfaces/ecpg/ecpglib/pg_type.h entirely, as the previous
commit reduced it to a confusingly-named wrapper around pg_type_d.h.
In passing, make test_rls_hooks.c follow project convention of including
our own files with #include "" not <>.
Discussion: https://postgr.es/m/23690.1523031777@sss.pgh.pa.us
pg_dump supposed that a stats object necessarily shares the same schema
as its underlying table, and that it doesn't have a separate owner.
These things may have been true during early development of the feature,
but they are not true as of v10 release.
Failure to track the object's schema separately turns out to have only
limited consequences, because pg_get_statisticsobjdef() always schema-
qualifies the target object name in the generated CREATE STATISTICS command
(a decision out of step with the rest of ruleutils.c, but I digress).
Therefore the restored object would be in the right schema, so that the
only problem is that the TOC entry would be mislabeled as to schema. That
could lead to wrong decisions for schema-selective restores, for example.
The ownership issue is a bit more serious: not only was the TOC entry
potentially mislabeled as to owner, but pg_dump didn't bother to issue an
ALTER OWNER command at all, so that after restore the stats object would
continue to be owned by the restoring superuser.
A final point is that decisions as to whether to dump a stats object or
not were driven by whether the underlying table was dumped or not. While
that's not wrong on its face, it won't scale nicely to the planned future
extension to cross-table statistics. Moreover, that design decision comes
out of the view of stats objects as being auxiliary to a particular table,
like a rule or trigger, which is exactly where the above problems came
from. Since we're now treating stats objects more like independent objects
in their own right, they ought to behave like standalone objects for this
purpose too. So change to using the generic selectDumpableObject() logic
for them (which presently amounts to "dump if containing schema is to be
dumped").
Along the way to fixing this, restructure so that getExtendedStatistics
collects the identity info (only) for all extended stats objects in one
query, and then for each object actually being dumped, we retrieve the
definition in dumpStatisticsExt. This is necessary to ensure that
schema-qualification in the generated CREATE STATISTICS command happens
with respect to the search path that pg_dump will now be using at restore
time (ie, the schema the stats object is in, not that of the underlying
table). It's probably also significantly faster in the typical scenario
where only a minority of tables have extended stats.
Back-patch to v10 where extended stats were introduced.
Discussion: https://postgr.es/m/18272.1518328606@sss.pgh.pa.us
When CREATE INDEX is run on a partitioned table, create catalog entries
for an index on the partitioned table (which is just a placeholder since
the table proper has no data of its own), and recurse to create actual
indexes on the existing partitions; create them in future partitions
also.
As a convenience gadget, if the new index definition matches some
existing index in partitions, these are picked up and used instead of
creating new ones. Whichever way these indexes come about, they become
attached to the index on the parent table and are dropped alongside it,
and cannot be dropped on isolation unless they are detached first.
To support pg_dump'ing these indexes, add commands
CREATE INDEX ON ONLY <table>
(which creates the index on the parent partitioned table, without
recursing) and
ALTER INDEX ATTACH PARTITION
(which is used after the indexes have been created individually on each
partition, to attach them to the parent index). These reconstruct prior
database state exactly.
Reviewed-by: (in alphabetical order) Peter Eisentraut, Robert Haas, Amit
Langote, Jesper Pedersen, Simon Riggs, David Rowley
Discussion: https://postgr.es/m/20171113170646.gzweigyrgg6pwsg4@alvherre.pgsql
This gets rid of the code that issued separate queries to retrieve the
partitioning parent-child relationship, parent partition key, and child
partition bound information. With this patch, the information is
retrieved instead using the queries issued from getTables() and
getInherits(), which is both more efficient than the previous approach
and doesn't require any new code.
Since the partitioning parent-child relationship is now retrieved with
the same old code that handles inheritance, partition attributes receive
a proper flagInhAttrs() treatment (that it didn't receive before), which
is needed so that the inherited NOT NULL constraints are not emitted if
we already emitted it for the parent.
Also, fix a bug in pg_dump's --binary-upgrade code, which caused pg_dump
to emit invalid command to attach a partition to its parent.
Author: Amit Langote, with some additional changes by me.
Add support for explicitly declared statistic objects (CREATE
STATISTICS), allowing collection of statistics on more complex
combinations that individual table columns. Companion commands DROP
STATISTICS and ALTER STATISTICS ... OWNER TO / SET SCHEMA / RENAME are
added too. All this DDL has been designed so that more statistic types
can be added later on, such as multivariate most-common-values and
multivariate histograms between columns of a single table, leaving room
for permitting columns on multiple tables, too, as well as expressions.
This commit only adds support for collection of n-distinct coefficient
on user-specified sets of columns in a single table. This is useful to
estimate number of distinct groups in GROUP BY and DISTINCT clauses;
estimation errors there can cause over-allocation of memory in hashed
aggregates, for instance, so it's a worthwhile problem to solve. A new
special pseudo-type pg_ndistinct is used.
(num-distinct estimation was deemed sufficiently useful by itself that
this is worthwhile even if no further statistic types are added
immediately; so much so that another version of essentially the same
functionality was submitted by Kyotaro Horiguchi:
https://postgr.es/m/20150828.173334.114731693.horiguchi.kyotaro@lab.ntt.co.jp
though this commit does not use that code.)
Author: Tomas Vondra. Some code rework by Álvaro.
Reviewed-by: Dean Rasheed, David Rowley, Kyotaro Horiguchi, Jeff Janes,
Ideriha Takeshi
Discussion: https://postgr.es/m/543AFA15.4080608@fuzzy.czhttps://postgr.es/m/20170320190220.ixlaueanxegqd5gr@alvherre.pgsql
Table partitioning is like table inheritance and reuses much of the
existing infrastructure, but there are some important differences.
The parent is called a partitioned table and is always empty; it may
not have indexes or non-inherited constraints, since those make no
sense for a relation with no data of its own. The children are called
partitions and contain all of the actual data. Each partition has an
implicit partitioning constraint. Multiple inheritance is not
allowed, and partitioning and inheritance can't be mixed. Partitions
can't have extra columns and may not allow nulls unless the parent
does. Tuples inserted into the parent are automatically routed to the
correct partition, so tuple-routing ON INSERT triggers are not needed.
Tuple routing isn't yet supported for partitions which are foreign
tables, and it doesn't handle updates that cross partition boundaries.
Currently, tables can be range-partitioned or list-partitioned. List
partitioning is limited to a single column, but range partitioning can
involve multiple columns. A partitioning "column" can be an
expression.
Because table partitioning is less general than table inheritance, it
is hoped that it will be easier to reason about properties of
partitions, and therefore that this will serve as a better foundation
for a variety of possible optimizations, including query planner
optimizations. The tuple routing based which this patch does based on
the implicit partitioning constraints is an example of this, but it
seems likely that many other useful optimizations are also possible.
Amit Langote, reviewed and tested by Robert Haas, Ashutosh Bapat,
Amit Kapila, Rajkumar Raghuwanshi, Corey Huinker, Jaime Casanova,
Rushabh Lathia, Erik Rijkers, among others. Minor revisions by me.
pg_dump has historically used a simple boolean 'dump' value to indicate
if a given object should be included in the dump or not. Instead, use
a bitmap which breaks down the components of an object into their
distinct pieces and use that bitmap to only include the components
requested.
This does not include any behavioral change, but is in preperation for
the change to dump out just ACLs for objects in pg_catalog.
Reviews by Alexander Korotkov, Jose Luis Tallon
Tom Lane
Bruce Momjian
Amit Kapila
Alvaro Herrera
Peter Geoghegan
Peter Eisentraut
Michael Paquier
Robert Haas
Stephen Frost
Peter Eisentraut