If a standby server has a cascading standby server connected to it, it's
possible that WAL has already been sent up to the next WAL page boundary,
splitting a WAL record in the middle, when the first standby server is
promoted. Don't throw an assertion failure or error in walsender if that
happens.
Also, fix a variant of the same bug in pg_receivexlog: if it had already
received WAL on previous timeline up to a segment boundary, when the
upstream standby server is promoted so that the timeline switch record falls
on the previous segment, pg_receivexlog would miss the segment containing
the timeline switch. To fix that, have walsender send the position of the
timeline switch at end-of-streaming, in addition to the next timeline's ID.
It was previously assumed that the switch happened exactly where the
streaming stopped.
Note: this is an incompatible change in the streaming protocol. You might
get an error if you try to stream over timeline switches, if the client is
running 9.3beta1 and the server is more recent. It should be fine after a
reconnect, however.
Tom Lane [Mon, 6 May 2013 17:26:51 +0000 (13:26 -0400)]
Move materialized views' is-populated status into their pg_class entries.
Previously this state was represented by whether the view's disk file had
zero or nonzero size, which is problematic for numerous reasons, since it's
breaking a fundamental assumption about heap storage. This was done to
allow unlogged matviews to revert to unpopulated status after a crash
despite our lack of any ability to update catalog entries post-crash.
However, this poses enough risk of future problems that it seems better to
not support unlogged matviews until we can find another way. Accordingly,
revert that choice as well as a number of existing kluges forced by it
in favor of creating a pg_class.relispopulated flag column.
Tom Lane [Mon, 6 May 2013 16:28:13 +0000 (12:28 -0400)]
Back out some recent translation updates.
Very old versions of msgfmt choke on these specific messages, for reasons
that are unclear at the moment. Remove them so that we can ship a beta
release and not get complaints from testers (these messages will just go
untranslated, instead, and we're hardly at 100% coverage anyway).
Peter Eisentraut will look for a better fix later.
Tom Lane [Mon, 6 May 2013 15:57:05 +0000 (11:57 -0400)]
Disallow unlogged materialized views.
The initial implementation of this feature was really unsupportable,
because it's relying on the physical size of an on-disk file to carry the
relation's populated/unpopulated state, which is at least a modularity
violation and could have serious long-term consequences. We could say that
an unlogged matview goes to empty on crash, but not everybody likes that
definition, so let's just remove the feature for 9.3. We can add it back
when we have a less klugy implementation.
I left the grammar and tab-completion support for CREATE UNLOGGED
MATERIALIZED VIEW in place, since it's harmless and allows delivering a
more specific error message about the unsupported feature.
I'm committing this separately to ease identification of what should be
reverted when/if we are able to re-enable the feature.
Tom Lane [Wed, 1 May 2013 22:26:50 +0000 (18:26 -0400)]
Fix permission tests for views/tables proven empty by constraint exclusion.
A view defined as "select <something> where false" had the curious property
that the system wouldn't check whether users had the privileges necessary
to select from it. More generally, permissions checks could be skipped
for tables referenced in sub-selects or views that were proven empty by
constraint exclusion (although some quick testing suggests this seldom
happens in cases of practical interest). This happened because the planner
failed to include rangetable entries for such tables in the finished plan.
This was noticed in connection with erroneous handling of materialized
views, but actually the issue is quite unrelated to matviews. Therefore,
revert commit 200ba1667b3a8d7a9d559d2f05f83d209c9d8267 in favor of a more
direct test for the real problem.
Simon Riggs [Tue, 30 Apr 2013 11:27:12 +0000 (12:27 +0100)]
Record data_checksum_version in control file.
The value is not used anywhere in code, but will
allow future changes to the checksum version
should that become necessary in the future.
Simon Riggs [Tue, 30 Apr 2013 07:15:49 +0000 (08:15 +0100)]
Ensure we MarkBufferDirty before visibilitymap_set()
logs the heap page and sets the LSN. Otherwise a
checkpoint could occur between those actions and
leave us in an inconsistent state.
Peter Eisentraut [Tue, 30 Apr 2013 02:17:29 +0000 (22:17 -0400)]
pg_upgrade: Remove PGPORT handling from test suite
This code was left over from when pg_upgrade paid attention to PGPORT.
Now it would only affects the regression test run before the test run of
pg_upgrade. You can still set PGPORT for that, but there is no reason
to have the test driver default it to 50432.
Peter Eisentraut [Tue, 30 Apr 2013 01:55:12 +0000 (21:55 -0400)]
Revert "pg_ctl: Add idempotent option"
This reverts commit 87306184580c9c49717b00d48a2f9e717f21e0a8. The
behavior in certain cases is still being debated, and it's too late to
solve this before beta.
Tom Lane [Mon, 29 Apr 2013 18:49:01 +0000 (14:49 -0400)]
Postpone creation of pathkeys lists to fix bug #8049.
This patch gets rid of the concept of, and infrastructure for,
non-canonical PathKeys; we now only ever create canonical pathkey lists.
The need for non-canonical pathkeys came from the desire to have
grouping_planner initialize query_pathkeys and related pathkey lists before
calling query_planner. However, since query_planner didn't actually *do*
anything with those lists before they'd been made canonical, we can get rid
of the whole mess by just not creating the lists at all until the point
where we formerly canonicalized them.
There are several ways in which we could implement that without making
query_planner itself deal with grouping/sorting features (which are
supposed to be the province of grouping_planner). I chose to add a
callback function to query_planner's API; other alternatives would have
required adding more fields to PlannerInfo, which while not bad in itself
would create an ABI break for planner-related plugins in the 9.2 release
series. This still breaks ABI for anything that calls query_planner
directly, but it seems somewhat unlikely that there are any such plugins.
I had originally conceived of this change as merely a step on the way to
fixing bug #8049 from Teun Hoogendoorn; but it turns out that this fixes
that bug all by itself, as per the added regression test. The reason is
that now get_eclass_for_sort_expr is adding the ORDER BY expression at the
end of EquivalenceClass creation not the start, and so anything that is in
a multi-member EquivalenceClass has already been created with correct
em_nullable_relids. I am suspicious that there are related scenarios in
which we still need to teach get_eclass_for_sort_expr to compute correct
nullable_relids, but am not eager to risk destabilizing either 9.2 or 9.3
to fix bugs that are only hypothetical. So for the moment, do this and
stop here.
Back-patch to 9.2 but not to earlier branches, since they don't exhibit
this bug for lack of join-clause-movement logic that depends on
em_nullable_relids being correct. (We might have to revisit that choice
if any related bugs turn up.) In 9.2, don't change the signature of
make_pathkeys_for_sortclauses nor remove canonicalize_pathkeys, so as
not to risk more plugin breakage than we have to.
Kevin Grittner [Mon, 29 Apr 2013 18:05:26 +0000 (13:05 -0500)]
Ensure ANALYZE phase is not skipped because of canceled truncate.
Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to
preserve existing behavior regarding statistics generation in the
case that a truncation attempt was canceled due to lock conflicts.
It failed to do this accurately in two regards: (1) autovacuum had
previously generated statistics if the truncate attempt failed to
initially get the lock rather than having started the attempt, and
(2) the VACUUM ANALYZE command had always generated statistics.
Both of these changes were unintended, and are reverted by this
patch. On review, there seems to be consensus that the previous
failure to generate statistics when the truncate was terminated
was more an unfortunate consequence of how that effort was
previously terminated than a feature we want to keep; so this
patch generates statistics even when an autovacuum truncation
attempt terminates early. Another unintended change which is kept
on the basis that it is an improvement is that when a VACUUM
command is truncating, it will the new heuristic for avoiding
blocking other processes, rather than keeping an
AccessExclusiveLock on the table for however long the truncation
takes.
Per multiple reports, with some renaming per patch by Jeff Janes.
Robert Haas [Mon, 29 Apr 2013 10:29:32 +0000 (06:29 -0400)]
Attempt to fix error recovery in COPY BOTH mode.
Previously, libpq and the backend had opposite ideas about whether
it was necessary for the client to send a CopyDone message after
receiving an ErrorResponse, making it impossible to cleanly exit
COPY BOTH mode. Fix libpq so that works correctly, adopting the
backend's notion that an ErrorResponse kills the copy in both
directions.
Adjust receivelog.c to avoid a degradation in the quality of the
resulting error messages. libpqwalreceiver.c is already doing
the right thing, so no adjustment needed there.
Add an explicit statement to the documentation explaining how
this part of the protocol is supposed to work, in the hopes of
avoiding future confusion in this area.
Since the consequences of all this confusion are very limited,
especially in the back-branches where no client ever attempts
to exit COPY BOTH mode without closing the connection entirely,
no back-patch.
Simon Riggs [Mon, 29 Apr 2013 08:05:27 +0000 (09:05 +0100)]
Introduce new page checksum algorithm and module.
Isolate checksum calculation to its own module, so that bufpage
knows little if anything about the details of the calculation.
This implementation is a modified FNV-1a hash checksum, details
of which are given in the new checksum.c header comments.
Basic implementation only, so we fix the output value.
Later related commits will add version numbers to pg_control,
compiler optimization flags and memory barriers.
Ants Aasma, reviewed by Jeff Davis and Simon Riggs
Tom Lane [Sun, 28 Apr 2013 04:18:45 +0000 (00:18 -0400)]
Editorialize a bit on new ProcessUtility() API.
Choose a saner ordering of parameters (adding a new input param after
the output params seemed a bit random), update the function's header
comment to match reality (cmon folks, is this really that hard?),
get rid of useless and sloppily-defined distinction between
PROCESS_UTILITY_SUBCOMMAND and PROCESS_UTILITY_GENERATED.
Tom Lane [Sun, 28 Apr 2013 03:11:28 +0000 (23:11 -0400)]
Fix unsafe event-trigger coding in ProcessUtility().
We mustn't run any of the event-trigger support code when handling
utility statements like START TRANSACTION or ABORT, because that code
may need to refresh event-trigger cache data, which requires being
inside a valid transaction. (This mistake explains the consistent
build failures exhibited by the CLOBBER_CACHE_ALWAYS buildfarm members,
as well as some irreproducible failures on other members.)
The least messy fix seems to be to break standard_ProcessUtility into two
functions, one that handles all the statements not supported by event
triggers, and one that contains the event-trigger support code and handles
the statements that are supported by event triggers.
This change also fixes several inconsistencies, such as four cases where
support had been installed for "ddl_event_start" but not "ddl_event_end"
triggers, plus the fact that InvokeDDLCommandEventTriggersIfSupported()
paid no mind to isCompleteQuery.
Tom Lane [Sat, 27 Apr 2013 21:48:57 +0000 (17:48 -0400)]
Incidental cleanup of matviews code.
Move checking for unscannable matviews into ExecOpenScanRelation, which is
a better place for it first because the open relation is already available
(saving a relcache lookup cycle), and second because this eliminates the
problem of telling the difference between rangetable entries that will or
will not be scanned by the query. In particular we can get rid of the
not-terribly-well-thought-out-or-implemented isResultRel field that the
initial matviews patch added to RangeTblEntry.
Also get rid of entirely unnecessary scannability check in the rewriter,
and a bogus decision about whether RefreshMatViewStmt requires a parse-time
snapshot.
catversion bump due to removal of a RangeTblEntry field, which changes
stored rules.
Tom Lane [Fri, 26 Apr 2013 19:48:24 +0000 (15:48 -0400)]
Fix collation assignment for aggregates with ORDER BY.
ORDER BY expressions were being treated the same as regular aggregate
arguments for purposes of collation determination, but really they should
not affect the aggregate's collation at all; only collations of the
aggregate's regular arguments should affect it.
In many cases this mistake would lead to incorrectly throwing a "collation
conflict" error; but in some cases the corrected code will silently assign
a different collation to the aggregate than before, for example
agg(foo ORDER BY bar COLLATE "x")
which will now use foo's collation rather than "x" for the aggregate.
Given this risk and the lack of field complaints about the issue, it
doesn't seem prudent to back-patch.
In passing, rearrange code in assign_collations_walker so that we don't
need multiple copies of the standard logic for computing collation of a
node with children. (Previously, CaseExpr duplicated the standard logic,
and we would have needed a third copy for Aggref without this change.)
Joe Conway [Fri, 26 Apr 2013 18:50:00 +0000 (11:50 -0700)]
Ensure that user created rows in extension tables get dumped if the table is explicitly requested, either with a -t/--table switch of the table itself, or by -n/--schema switch of the schema containing the extension table. Patch reviewed by Vibhor Kumar and Dimitri Fontaine.
Backpatched to 9.1 when the extension management facility was added.
Robert Haas [Fri, 26 Apr 2013 12:57:47 +0000 (08:57 -0400)]
libpq: Fix a few bits that didn't get the memo about COPY BOTH.
There's probably no real bug here at present, so not backpatching.
But it seems good to make these bits consistent with the rest of
libpq, so as to avoid future surprises.
Tom Lane [Thu, 25 Apr 2013 20:58:05 +0000 (16:58 -0400)]
Avoid deadlock between concurrent CREATE INDEX CONCURRENTLY commands.
There was a high probability of two or more concurrent C.I.C. commands
deadlocking just before completion, because each would wait for the others
to release their reference snapshots. Fix by releasing the snapshot
before waiting for other snapshots to go away.
Per report from Paul Hinze. Back-patch to all active branches.
On non-Windows systems, sys/time.h was pulled in by portability/instr_time.h,
which pulled in time.h. We certainly should include time.h directly, since
we're using time(2), but the indirect include masked the problem on most
platforms.
Kevin Grittner [Wed, 24 Apr 2013 13:39:06 +0000 (08:39 -0500)]
Fix assertion failure for REFRESH MATERIALIZED VIEW in PL.
This was due to incomplete implementation of rowcount reporting
for RMV, which was due to initial waffling on whether it should
be provided. It seems unlikely to be a useful or universally
available number as more sophisticated techniques for maintaining
matviews are added, so remove the partial support rather than
completing it.
Per report of Jeevan Chalke, but with a different fix
Tom Lane [Sat, 20 Apr 2013 20:59:21 +0000 (16:59 -0400)]
Fix longstanding race condition in plancache.c.
When creating or manipulating a cached plan for a transaction control
command (particularly ROLLBACK), we must not perform any catalog accesses,
since we might be in an aborted transaction. However, plancache.c busily
saved or examined the search_path for every cached plan. If we were
unlucky enough to do this at a moment where the path's expansion into
schema OIDs wasn't already cached, we'd do some catalog accesses; and with
some more bad luck such as an ill-timed signal arrival, that could lead to
crashes or Assert failures, as exhibited in bug #8095 from Nachiket Vaidya.
Fortunately, there's no real need to consider the search path for such
commands, so we can just skip the relevant steps when the subject statement
is a TransactionStmt. This is somewhat related to bug #5269, though the
failure happens during initial cached-plan creation rather than
revalidation.
This bug has been there since the plan cache was invented, so back-patch
to all supported branches.
Peter Eisentraut [Sat, 20 Apr 2013 15:04:41 +0000 (11:04 -0400)]
Clean up references to SQL92
In most cases, these were just references to the SQL standard in
general. In a few cases, a contrast was made between SQL92 and later
standards -- those have been kept unchanged.
Tom Lane [Fri, 19 Apr 2013 20:14:56 +0000 (16:14 -0400)]
Improve error message when an FDW doesn't support WHERE CURRENT OF.
If an FDW fails to take special measures with a CurrentOfExpr, we will
end up trying to execute it as an ordinary qual, which was being treated
as a purely internal failure condition. Provide a more user-oriented
error message for such cases.
Remove some unused and seldom used fields from RelationAmInfo.
This saves some memory from each index relcache entry. At least on a 64-bit
machine, it saves just enough to shrink a typical relcache entry's memory
usage from 2k to 1k. That's nice if you have a lot of backends and a lot of
indexes.