f40d362a66 disabled part of 031_recovery_conflict.pl due to instability
that's not trivial to fix in the back branches. That fixed most of the
issues. But there was one more failure (on lapwing / REL_10_STABLE).
That failure looks like it might be caused by a genuine problem. Disable the
test until after the set of releases, to avoid packagers etc potentially
having to fight with a test failure they can't do anything about.
Discussion: https://postgr.es/m/3447060.1652032749@sss.pgh.pa.us
Backpatch: 10-14
The recovery deadlock test has a timing issue that was fixed in 5136967f1e in
HEAD. Unfortunately the same fix doesn't quite work in the back branches: 1)
adjust_conf() doesn't exist, which is easy enough to work around 2) a restart
cleares the recovery conflict stats < 15.
These issues can be worked around, but given the upcoming set of minor
releases, skip the problematic test for now. The buildfarm doesn't show
failures in other parts of 031_recovery_conflict.pl.
Discussion: https://postgr.es/m/20220506155827.dfnaheq6ufylwrqf@alap3.anarazel.de
Backpatch: 10-14
The tests added in 9f8a050f68 failed nearly reliably on FreeBSD in CI, and
occasionally on the buildfarm. That turns out to be caused not by a bug in the
test, but by a longstanding bug in recovery conflict handling.
The standby timeout handler, used by ResolveRecoveryConflictWithBufferPin(),
executed SendRecoveryConflictWithBufferPin() inside a signal handler. A bad
idea, because the deadlock timeout handler (or a spurious latch set) could
have interrupted ProcWaitForSignal(). If unlucky that could cause a
self-deadlock on ProcArrayLock, if the deadlock check is in
SendRecoveryConflictWithBufferPin()->CancelDBBackends().
To fix, set a flag in StandbyTimeoutHandler(), and check the flag in
ResolveRecoveryConflictWithBufferPin().
Subsequently the recovery conflict tests will be backpatched.
Discussion: https://postgr.es/m/20220413002626.udl7lll7f3o7nre7@alap3.anarazel.de
Backpatch: 10-
mark_async_capable_plan(), which is called from create_append_plan() to
determine whether subplans are async-capable, failed to take into
account that the given subplan created from a given subpath might
include a gating Result node if the subpath is a SubqueryScanPath or
ForeignPath, causing a segmentation fault there when the subplan created
from a SubqueryScanPath includes the Result node, or causing
ExecAsyncRequest() to throw an error about an unrecognized node type
when the subplan created from a ForeignPath includes the Result node,
because in the latter case the Result node was unintentionally
considered as async-capable, but we don't currently support executing
Result nodes asynchronously. Fix by modifying mark_async_capable_plan()
to disable asynchronous execution in such cases. Also, adjust code in
the ProjectionPath case in mark_async_capable_plan(), for consistency
with other cases, and adjust/improve comments there.
is_async_capable_path() added in commit 27e1f1456, which was rewritten
to mark_async_capable_plan() in a later commit, has the same issue,
causing the error at execution mentioned above, so back-patch to v14
where the aforesaid commit went in.
Per report from Justin Pryzby.
Etsuro Fujita, reviewed by Zhihong Yu and Justin Pryzby.
Discussion: https://postgr.es/m/20220408124338.GK24419%40telsasoft.com
For some reason by default the mingw C Runtime takes it upon itself to
expand program arguments that look like shell globbing characters. That
has caused much scratching of heads and mis-attribution of the causes of
some TAP test failures, so stop doing that.
This removes an inconsistency with Windows binaries built with MSVC,
which have no such behaviour.
Per suggestion from Noah Misch.
Backpatch to all live branches.
Discussion: https://postgr.es/m/20220423025927.GA1274057@rfd.leadboat.com
This function looks for a reference to the recursive WITH CTE,
but it checked only the CTE name not ctelevelsup, so that it could
seize on a lower CTE that happened to have the same name. This
would result in planner failures later, either weird errors such as
"could not find attribute 2 in subquery targetlist", or crashes
or assertion failures. The code also merely Assert'ed that it found
a matching entry, which is not guaranteed at all by the parser.
Per bugs #17320 and #17318 from Zhiyong Wu.
Thanks to Kyotaro Horiguchi for investigation.
Discussion: https://postgr.es/m/17320-70e37868182512ab@postgresql.org
Discussion: https://postgr.es/m/17318-2eb65a3a611d2368@postgresql.org
inline_cte() expected to find exactly as many references to the
target CTE as its cterefcount indicates. While that should be
accurate for the tree as emitted by the parser, there are some
optimizations that occur upstream of here that could falsify it,
notably removal of unused subquery output expressions.
Trying to make the accounting 100% accurate seems expensive and
doomed to future breakage. It's not really worth it, because
all this code is protecting is downstream assumptions that every
referenced CTE has a plan. Let's convert those assertions to
regular test-and-elog just in case there's some actual problem,
and then drop the failing assertion.
Per report from Tomas Vondra (thanks also to Richard Guo for
analysis). Back-patch to v12 where the faulty code came in.
Discussion: https://postgr.es/m/29196a1e-ed47-c7ca-9be2-b1c636816183@enterprisedb.com
Commit b3b4d8e68a moved our perl test modules to a better namespace
structure, but this has made life hard for people wishing to backpatch
improvements in the TAP tests. Here we alleviate much of that difficulty
by implementing the new module names on top of the old modules, mostly
by using a little perl typeglob aliasing magic, so that we don't have a
dual maintenance burden. This should work both for the case where a new
test is backpatched and the case where a fix to an existing test that
uses the new namespace is backpatched.
Reviewed by Michael Paquier
Per complaint from Andres Freund
Discussion: https://postgr.es/m/20220418141530.nfxtkohefvwnzncl@alap3.anarazel.de
Applied to branches 10 through 14
CLUSTER sort won't use the datum1 SortTuple field when clustering
against an index whose leading key is an expression. This makes it
unsafe to use the abbreviated keys optimization, which was missed by the
logic that sets up SortSupport state. Affected tuplesorts output tuples
in a completely bogus order as a result (the wrong SortSupport based
comparator was used for the leading attribute).
This issue is similar to the bug fixed on the master branch by recent
commit cc58eecc5d. But it's a far older issue, that dates back to the
introduction of the abbreviated keys optimization by commit 4ea51cdfe8.
Backpatch to all supported versions.
Author: Peter Geoghegan <pg@bowt.ie>
Author: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/CA+hUKG+bA+bmwD36_oDxAoLrCwZjVtST2fqe=b4=qZcmU7u89A@mail.gmail.com
Backpatch: 10-
Such cases will lead to infinite loops, so they're of no practical
value. The numeric variant of generate_series() already threw error
for this, so borrow its message wording.
Per report from Richard Wesley. Back-patch to all supported branches.
Discussion: https://postgr.es/m/91B44E7B-68D5-448F-95C8-B4B3B0F5DEAF@duckdblabs.com
psql, pg_dump, and pg_amcheck share code to process object name
patterns like 'foo*.bar*' to match all tables with names starting in
'bar' that are in schemas starting with 'foo'. Before v14, any number
of extra name parts were silently ignored, so a command line '\d
foo.bar.baz.bletch.quux' was interpreted as '\d bletch.quux'. In v14,
as a result of commit 2c8726c4b0, we
instead treated this as a request for table quux in a schema named
'foo.bar.baz.bletch'. That caused problems for people like Justin
Pryzby who were accustomed to copying strings of the form
db.schema.table from messages generated by PostgreSQL itself and using
them as arguments to \d.
Accordingly, revise things so that if an object name pattern contains
more parts than we're expecting, we throw an error, unless there's
exactly one extra part and it matches the current database name.
That way, thisdb.myschema.mytable is accepted as meaning just
myschema.mytable, but otherdb.myschema.mytable is an error, and so
is some.random.garbage.myschema.mytable.
Mark Dilger, per report from Justin Pryzby and discussion among
various people.
Discussion: https://www.postgresql.org/message-id/20211013165426.GD27491%40telsasoft.com
An ALTER FUNCTION command that tried to update both the function's
proparallel property and its proconfig list failed to do the former,
because it stored the new proparallel value into a tuple that was
no longer the interesting one. Carelessness in 7aea8e4f2.
(I did not bother with a regression test, because the only likely
future breakage would be for someone to ignore the comment I added
and add some other field update after the heap_modify_tuple step.
A test using existing function properties could not catch that.)
Per report from Bryn Llewellyn. Back-patch to all supported branches.
Discussion: https://postgr.es/m/8AC9A37F-99BD-446F-A2F7-B89AD0022774@yugabyte.com
Don't try to look at the attidentity field of system attributes,
because they're not there in the TupleDescAttr array. Sometimes
this is harmless because we accidentally pick up a zero, but
otherwise we'll report "no owned sequence found" from an attempt
to alter a system attribute. (It seems possible that a SIGSEGV
could occur, too, though I've not seen it in testing.)
It's not in this function's charter to complain that you can't
alter a system column, so instead just hard-wire an assumption
that system attributes aren't identities. I didn't bother with
a regression test because the appearance of the bug is very
erratic.
Per bug #17465 from Roman Zharkov. Back-patch to all supported
branches. (There's not actually a live bug before v12, because
before that get_attidentity() did the right thing anyway.
But for consistency I changed the test in the older branches too.)
Discussion: https://postgr.es/m/17465-f2a554a6cb5740d3@postgresql.org
This test, introduced in df86e52, uses a second standby to check that
it is able to remove correctly RECOVERYHISTORY and RECOVERYXLOG at the
end of recovery. This standby uses the archives of the primary to
restore its contents, with some of the archive's contents coming from
the first standby previously promoted. In slow environments, it was
possible that the test did not check what it should, as the history file
generated by the promotion of the first standby may not be stored yet on
the archives the second standby feeds on. So, it could be possible that
the second standby selects an incorrect timeline, without restoring a
history file at all.
This commits adds a wait phase to make sure that the history file
required by the second standby is archived before this cluster is
created. This relies on poll_query_until() with pg_stat_file() and an
absolute path, something not supported in REL_10_STABLE.
While on it, this adds a new test to check that the history file has
been restored by looking at the logs of the second standby. This
ensures that a RECOVERYHISTORY, whose removal needs to be checked,
is created in the first place. This should make the test more robust.
This test has been introduced by df86e52, but it came in light as an
effect of the bug fixed by acf1dd42, where the extra restore_command
calls made the test much slower.
Reported-by: Andres Freund
Discussion: https://postgr.es/m/YlT23IvsXkGuLzFi@paquier.xyz
Backpatch-through: 11
The target failed, tested $PATH binaries, or tested a stale temporary
installation. Commit c66b438db6 missed
this. Back-patch to v10 (all supported versions).
The back-patch of commit bbace5697d had
the unfortunate effect of changing the layout of PGPROC in the
back-branches, which could break extensions. This happened because it
changed the delayChkpt from type bool to type int. So, change it back,
and add a new bool delayChkptEnd field instead. The new field should
fall within what used to be padding space within the struct, and so
hopefully won't cause any extensions to break.
Per report from Markus Wanner and discussion with Tom Lane and others.
Patch originally by me, somewhat revised by Markus Wanner per a
suggestion from Michael Paquier. A very similar patch was developed
by Kyotaro Horiguchi, but I failed to see the email in which that was
posted before writing one of my own.
Discussion: http://postgr.es/m/CA+Tgmoao-kUD9c5nG5sub3F7tbo39+cdr8jKaOVEs_1aBWcJ3Q@mail.gmail.com
Discussion: http://postgr.es/m/20220406.164521.17171257901083417.horikyota.ntt@gmail.com
heap_fetch() used to have a "keep_buf" parameter that told it to return
ownership of the buffer pin to the caller after finding that the
requested tuple TID exists but is invisible to the specified snapshot.
This was thoughtlessly removed in commit 5db6df0c0, which broke
heapam_tuple_lock() (formerly EvalPlanQualFetch) because that function
needs to do more accesses to the tuple even if it's invisible. The net
effect is that we would continue to touch the page for a microsecond or
two after releasing pin on the buffer. Usually no harm would result;
but if a different session decided to defragment the page concurrently,
we could see garbage data and mistakenly conclude that there's no newer
tuple version to chain up to. (It's hard to say whether this has
happened in the field. The bug was actually found thanks to a later
change that allowed valgrind to detect accesses to non-pinned buffers.)
The most reasonable way to fix this is to reintroduce keep_buf,
although I made it behave slightly differently: buffer ownership
is passed back only if there is a valid tuple at the requested TID.
In HEAD, we can just add the parameter back to heap_fetch().
To avoid an API break in the back branches, introduce an additional
function heap_fetch_extended() in those branches.
In HEAD there is an additional, less obvious API change: tuple->t_data
will be set to NULL in all cases where buffer ownership is not returned,
in particular when the tuple exists but fails the time qual (and
!keep_buf). This is to defend against any other callers attempting to
access non-pinned buffers. We concluded that making that change in back
branches would be more likely to introduce problems than cure any.
In passing, remove a comment about heap_fetch that was obsoleted by
9a8ee1dc6.
Per bug #17462 from Daniil Anisimov. Back-patch to v12 where the bug
was introduced.
Discussion: https://postgr.es/m/17462-9c98a0f00df9bd36@postgresql.org
It's possible for the query that "waits for restart" to complete a
successful iteration before the postmaster has noticed its SIGKILL'd
child and begun the restart cycle. (This is a bit hard to believe
perhaps, but it's been seen at least twice in the buildfarm, mainly
on ancient platforms that likely have quirky schedulers.)
To provide a more secure interlock, wait for the other session
we're using to report that it's been forcibly shut down.
Patch by me, based on a suggestion from Andres Freund.
Back-patch to v14 where this test case came in.
Discussion: https://postgr.es/m/1801850.1649047827@sss.pgh.pa.us
The expected backend message after SIGQUIT changed in commit
7e784d1dc, but we missed updating this test case. Also, experience
shows that we might sometimes get "could not send data to server"
instead of either of the libpq messages the test is looking for.
Per report from Mark Dilger. Back-patch to v14 where the
backend message changed.
Discussion: https://postgr.es/m/17BD82D7-49AC-40C9-8204-E7ADD30321A0@enterprisedb.com
pg_stat_get_replication_slot() accidentally was marked as non-strict, crashing
when called with NULL input. As it's already released, introduce an explicit
NULL check in 14, fix the catalog in HEAD.
Bumps catversion in HEAD.
Discussion: https://postgr.es/m/20220326212432.s5n2maw6kugnpyxw@alap3.anarazel.de
Backpatch: 14-, where replication slot stats were introduced
After closedir() dirent->d_name is not valid anymore. As there alerady are a
few places relying on the limited lifetime of pg_waldump, do so here as well,
and just pg_strdup() the string.
The bug was introduced in fc49e24fa6.
Found by UBSan, run locally.
Backpatch: 11-, like fc49e24fa6 itself.
Commit 8c6d30f21 caused this function to fail to set *displen
in the PS_USE_NONE code path. If the variable's previous value
had been negative, that'd lead to a memory clobber at some call
sites. We'd managed not to notice due to very thin test coverage
of such configurations, but this appears to explain buildfarm member
lorikeet's recent struggles.
Credit to Andrew Dunstan for spotting the problem. Back-patch
to v13 where the bug was introduced.
Discussion: https://postgr.es/m/136102.1648320427@sss.pgh.pa.us
clang 13 with -Wextra warns that "performing pointer subtraction with
a null pointer has undefined behavior" in the places where freepage.c
tries to set a relptr variable to constant NULL. This appears to be
a compiler bug, but it's unlikely to get fixed instantly. Fortunately,
we can work around it by introducing an inline support function, which
seems like a good change anyway because it removes the macro's existing
double-evaluation hazard.
Backpatch to v10 where this code was introduced.
Patch by me, based on an idea of Andres Freund's.
Discussion: https://postgr.es/m/48826.1648310694@sss.pgh.pa.us
The previous method for doing that was to write zeroes into a
predetermined set of page locations. However, there's a roughly
1-in-64K chance that the existing checksum will match by chance,
and yesterday several buildfarm animals started to reproducibly
see that, resulting in test failures because no checksum mismatch
was reported.
Since the checksum includes the page LSN, test success depends on
the length of the installation's WAL history, which is affected by
(at least) the initial catalog contents, the set of locales installed
on the system, and the length of the pathname of the test directory.
Sooner or later we were going to hit a chance match, and today is
that day.
Harden these tests by specifically inverting the checksum field and
leaving all else alone, thereby guaranteeing that the checksum is
incorrect.
In passing, fix places that were using seek() to set up for syswrite(),
a combination that the Perl docs very explicitly warn against. We've
probably escaped problems because no regular buffered I/O is done on
these filehandles; but if it ever breaks, we wouldn't deserve or get
much sympathy.
Although we've only seen problems in HEAD, now that we recognize the
environmental dependencies it seems like it might be just a matter
of time until someone manages to hit this in back-branch testing.
Hence, back-patch to v11 where we started doing this kind of test.
Discussion: https://postgr.es/m/3192026.1648185780@sss.pgh.pa.us
Crash recovery on standby may encounter missing directories when
replaying create database WAL records. Prior to this patch, the standby
would fail to recover in such a case. However, the directories could be
legitimately missing. Consider a sequence of WAL records as follows:
CREATE DATABASE
DROP DATABASE
DROP TABLESPACE
If, after replaying the last WAL record and removing the tablespace
directory, the standby crashes and has to replay the create database
record again, the crash recovery must be able to move on.
This patch adds a mechanism similar to invalid-page tracking, to keep a
tally of missing directories during crash recovery. If all the missing
directory references are matched with corresponding drop records at the
end of crash recovery, the standby can safely continue following the
primary.
Backpatch to 13, at least for now. The bug is older, but fixing it in
older branches requires more careful study of the interactions with
commit e6d8069522, which appeared in 13.
A new TAP test file is added to verify the condition. However, because
it depends on commit d6d317dbf6, it can only be added to branch
master. I (Álvaro) manually verified that the code behaves as expected
in branch 14. It's a bit nervous-making to leave the code uncovered by
tests in older branches, but leaving the bug unfixed is even worse.
Also, the main reason this fix took so long is precisely that we
couldn't agree on a good strategy to approach testing for the bug, so
perhaps this is the best we can do.
Diagnosed-by: Paul Guo <paulguo@gmail.com>
Author: Paul Guo <paulguo@gmail.com>
Author: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Author: Asim R Praveen <apraveen@pivotal.io>
Discussion: https://postgr.es/m/CAEET0ZGx9AvioViLf7nbR_8tH9-=27DN5xWJ2P9-ROH16e4JUA@mail.gmail.com
If TRUNCATE causes some buffers to be invalidated and thus the
checkpoint does not flush them, TRUNCATE must also ensure that the
corresponding files are truncated on disk. Otherwise, a replay
from the checkpoint might find that the buffers exist but have
the wrong contents, which may cause replay to fail.
Report by Teja Mupparti. Patch by Kyotaro Horiguchi, per a design
suggestion from Heikki Linnakangas, with some changes to the
comments by me. Review of this and a prior patch that approached
the issue differently by Heikki Linnakangas, Andres Freund, Álvaro
Herrera, Masahiko Sawada, and Tom Lane.
Discussion: http://postgr.es/m/BYAPR06MB6373BF50B469CA393C614257ABF00@BYAPR06MB6373.namprd06.prod.outlook.com
Invalidate abortedRecPtr and missingContrecPtr after a missing
continuation record is successfully skipped on a standby. This fixes a
PANIC caused when a recently promoted standby attempts to write an
OVERWRITE_RECORD with an LSN of the previously read aborted record.
Backpatch to 10 (all stable versions).
Author: Sami Imseih <simseih@amazon.com>
Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Reviewed-by: Álvaro Herrera <alvherre@alvh.no-ip.org>
Discussion: https://postgr.es/m/44D259DE-7542-49C4-8A52-2AB01534DCA9@amazon.com
When cross-building to windows, or building with mingw on windows, the build
could fail with
x86_64-w64-mingw32-gcc: error: win32ver.o: No such file or director
because pg_dumpall didn't depend on WIN32RES, but it's recipe references
it. The build nevertheless succeeded most of the time, due to
pg_dump/pg_restore having the required dependency, causing win32ver.o to be
built.
Reported-By: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/CA+hUKGJeekpUPWW6yCVdf9=oBAcCp86RrBivo4Y4cwazAzGPng@mail.gmail.com
Backpatch: 10-, omission present on all live branches
This issue is environment-sensitive, where the SSL tests could fail in
various way by feeding on defaults provided by sslcert, sslkey,
sslrootkey, sslrootcert, sslcrl and sslcrldir coming from a local setup,
as of ~/.postgresql/ by default. Horiguchi-san has reported two
failures, but more advanced testing from me (aka inclusion of garbage
SSL configuration in ~/.postgresql/ for all the configuration
parameters) has showed dozens of failures that can be triggered in the
whole test suite.
History has showed that we are not good when it comes to address such
issues, fixing them locally like in dd87799, and such problems keep
appearing. This commit strengthens the entire test suite to put an end
to this set of problems by embedding invalid default values in all the
connection strings used in the tests. The invalid values are prefixed
in each connection string, relying on the follow-up values passed in the
connection string to enforce any invalid value previously set. Note
that two tests related to CRLs are required to fail with certain pre-set
configurations, but we can rely on enforcing an empty value instead
after the invalid set of values.
Reported-by: Kyotaro Horiguchi
Reviewed-by: Andrew Dunstan, Daniel Gustafsson, Kyotaro Horiguchi
Discussion: https://postgr.es/m/20220316.163658.1122740600489097632.horikyota.ntt@gmail.com
backpatch-through: 10
