When building with sanitizers the sanitizer library provides dlopen, but not
dlsym(), making configure think that -ldl isn't needed. Just checking for
dlsym() ought to suffice, hard to see dlsym() being provided without dlopen()
also being provided.
Backpatch to all branches, for the same reasons as 46ab07ffda.
Reviewed-By: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/20220323173537.ll7klrglnp4gn2um@alap3.anarazel.de
Backpatch: 10-
It seems possible for the condition being tested to be true in
production, and nobody would never know (except when some data
eventually becomes corrupt?).
Author: Álvaro Herrera <alvherre@alvh.no-ip.org>
Discussion: https://postgr.es/m//202109040001.zky3wgv2qeqg@alvherre.pgsql
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
The planner needs to treat GroupingFunc like Aggref for many purposes,
in particular with respect to processing of the argument expressions,
which are not to be evaluated at runtime. A few places hadn't gotten
that memo, notably including subselect.c's processing of outer-level
aggregates. This resulted in assertion failures or wrong plans for
cases in which a GROUPING() construct references an outer aggregation
level.
Also fix missing special cases for GroupingFunc in cost_qual_eval
(resulting in wrong cost estimates for GROUPING(), although it's
not clear that that would affect plan shapes in practice) and in
ruleutils.c (resulting in excess parentheses in pretty-print mode).
Per bug #17088 from Yaoguang Chen. Back-patch to all supported
branches.
Richard Guo, Tom Lane
Discussion: https://postgr.es/m/17088-e33882b387de7f5c@postgresql.org
DROP INDEX needs to lock the index's table before the index itself,
else it will deadlock against ordinary queries that acquire the
relation locks in that order. This is correctly mechanized for
plain indexes by RangeVarCallbackForDropRelation; but in the case of
a partitioned index, we neglected to lock the child tables in advance
of locking the child indexes. We can fix that by traversing the
inheritance tree and acquiring the needed locks in RemoveRelations,
after we have acquired our locks on the parent partitioned table and
index.
While at it, do some refactoring to eliminate confusion between
the actual and expected relkind in RangeVarCallbackForDropRelation.
We can save a couple of syscache lookups too, by having that function
pass back info that RemoveRelations will need.
Back-patch to v11 where partitioned indexes were added.
Jimmy Yih, Gaurab Dey, Tom Lane
Discussion: https://postgr.es/m/BYAPR05MB645402330042E17D91A70C12BD5F9@BYAPR05MB6454.namprd05.prod.outlook.com
The output of table_to_xmlschema() and allied functions includes
a regex describing valid values for these types ... but the regex
was itself invalid, as it failed to escape a literal "+" sign.
Report and fix by Renan Soares Lopes. Back-patch to all
supported branches.
Discussion: https://postgr.es/m/7f6fabaa-3f8f-49ab-89ca-59fbfe633105@me.com
In commit bf7ca1587, I had the bright idea that we could make the
result of a whole-row Var (that is, foo.*) track any column aliases
that had been applied to the FROM entry the Var refers to. However,
that's not terribly logically consistent, because now the output of
the Var is no longer of the named composite type that the Var claims
to emit. bf7ca1587 tried to handle that by changing the output
tuple values to be labeled with a blessed RECORD type, but that's
really pretty disastrous: we can wind up storing such tuples onto
disk, whereupon they're not readable by other sessions.
The only practical fix I can see is to give up on what bf7ca1587
tried to do, and say that the column names of tuples produced by
a whole-row Var are always those of the underlying named composite
type, query aliases or no. While this introduces some inconsistencies,
it removes others, so it's not that awful in the abstract. What *is*
kind of awful is to make such a behavioral change in a back-patched
bug fix. But corrupt data is worse, so back-patched it will be.
(A workaround available to anyone who's unhappy about this is to
introduce an extra level of sub-SELECT, so that the whole-row Var is
referring to the sub-SELECT's output and not to a named table type.
Then the Var is of type RECORD to begin with and there's no issue.)
Per report from Miles Delahunty. The faulty commit dates to 9.5,
so back-patch to all supported branches.
Discussion: https://postgr.es/m/2950001.1638729947@sss.pgh.pa.us
Commands like ALTER TABLE SET TABLESPACE may leave files for the next
checkpoint to clean up. If such files are not removed by the time DROP
TABLESPACE is called, we request a checkpoint so that they are deleted.
However, there is presently a window before checkpoint start where new
unlink requests won't be scheduled until the following checkpoint. This
means that the checkpoint forced by DROP TABLESPACE might not remove the
files we expect it to remove, and the following ERROR will be emitted:
ERROR: tablespace "mytblspc" is not empty
To fix, add a call to AbsorbSyncRequests() just before advancing the
unlink cycle counter. This ensures that any unlink requests forwarded
prior to checkpoint start (i.e., when ckpt_started is incremented) will
be processed by the current checkpoint. Since AbsorbSyncRequests()
performs memory allocations, it cannot be called within a critical
section, so we also need to move SyncPreCheckpoint() to before
CreateCheckPoint()'s critical section.
This is an old bug, so back-patch to all supported versions.
Author: Nathan Bossart <nathandbossart@gmail.com>
Reported-by: Nathan Bossart <nathandbossart@gmail.com>
Reviewed-by: Thomas Munro <thomas.munro@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/20220215235845.GA2665318%40nathanxps13
This caused the function to fail, as the aligned copy of the raw page
given by the function caller was not saved in the correct memory
context, which needs to be multi_call_memory_ctx in this case.
Issue introduced by 076f4d9.
Per buildfarm members sifika, mylodon and longfin. I have reproduced
that locally with macos.
Discussion: https://postgr.es/m/YjFPOtfCW6yLXUeM@paquier.xyz
Backpatch-through: 10
This commit fixes a set of issues related to the use of the SQL
functions in this module when the caller is able to pass down raw page
data as input argument:
- The page size check was fuzzy in a couple of places, sometimes
looking after only a sub-range, but what we are looking for is an exact
match on BLCKSZ. After considering a few options here, I have settled
down to do a generalization of get_page_from_raw(). Most of the SQL
functions already used that, and this is not strictly required if not
accessing an 8-byte-wide value from a raw page, but this feels safer in
the long run for alignment-picky environment, particularly if a code
path begins to access such values. This also reduces the number of
strings that need to be translated.
- The BRIN function brin_page_items() uses a Relation but it did not
check the access method of the opened index, potentially leading to
crashes. All the other functions in need of a Relation already did
that.
- Some code paths could fail on elog(), but we should to use ereport()
for failures that can be triggered by the user.
Tests are added to stress all the cases that are fixed as of this
commit, with some junk raw pages (\set VERBOSITY ensures that this works
across all page sizes) and unexpected index types when functions open
relations.
Author: Michael Paquier, Justin Prysby
Discussion: https://postgr.es/m/20220218030020.GA1137@telsasoft.com
Backpatch-through: 10
Since LLVM 14 has stopped changing and is about to be released,
back-patch the following changes from the master branch:
e6a7600202807fee1a39a56e7b6601
Back-patch to 11, where LLVM JIT support came in.
ALTER ROUTINE triggers the events ddl_command_start and ddl_command_end,
and DROP ROUTINE triggers sql_drop, ddl_command_start and
ddl_command_end, but this was not mention on the matrix table.
Reported-by: Leslie Lemaire
Discussion: https://postgr.es/m/164647533363.646.5802968483136493025@wrigleys.postgresql.org
Backpatch-through: 11
Slow hosts may avoid load-induced, spurious failures by setting
environment variable PG_TEST_TIMEOUT_DEFAULT to some number of seconds
greater than 180. Developers may see faster failures by setting that
environment variable to some lesser number of seconds. In tests, write
$PostgreSQL::Test::Utils::timeout_default wherever the convention has
been to write 180. This change raises the default for some briefer
timeouts. Back-patch to v10 (all supported versions).
Discussion: https://postgr.es/m/20220218052842.GA3627003@rfd.leadboat.com
This macro cast the result to BlockNumber after shifting, not before,
which is the wrong thing. Per the C spec, the uint16 fields would
promote to int not unsigned int, so that (for 32-bit int) the shift
potentially shifts a nonzero bit into the sign position. I doubt
there are any production systems where this would actually end with
the wrong answer, but it is undefined behavior per the C spec, and
clang's -fsanitize=undefined option reputedly warns about it on some
platforms. (I can't reproduce that right now, but the code is
undeniably wrong per spec.) It's easy to fix by casting to
BlockNumber (uint32) in the proper places.
It's been wrong for ages, so back-patch to all supported branches.
Report and patch by Zhihong Yu (cosmetic tweaking by me)
Discussion: https://postgr.es/m/CALNJ-vT9r0DSsAOw9OXVJFxLENoVS_68kJ5x0p44atoYH+H4dg@mail.gmail.com
Most of these are cases where we could call memcpy() or other libc
functions with a NULL pointer and a zero count, which is forbidden
by POSIX even though every production version of libc allows it.
We've fixed such things before in a piecemeal way, but apparently
never made an effort to try to get them all. I don't claim that
this patch does so either, but it gets every failure I observe in
check-world, using clang 12.0.1 on current RHEL8.
numeric.c has a different issue that the sanitizer doesn't like:
"ln(-1.0)" will compute log10(0) and then try to assign the
resulting -Inf to an integer variable. We don't actually use the
result in such a case, so there's no live bug.
Back-patch to all supported branches, with the idea that we might
start running a buildfarm member that tests this case. This includes
back-patching c1132aae3 (Check the size in COPY_POINTER_FIELD),
which previously silenced some of these issues in copyfuncs.c.
Discussion: https://postgr.es/m/CALNJ-vT9r0DSsAOw9OXVJFxLENoVS_68kJ5x0p44atoYH+H4dg@mail.gmail.com
This change makes libpq apply the same private-key-file ownership
and permissions checks that we have used in the backend since commit
9a83564c5. Namely, that the private key can be owned by either the
current user or root (with different file permissions allowed in the
two cases). This allows system-wide management of key files, which
is just as sensible on the client side as the server, particularly
when the client is itself some application daemon.
Sync the comments about this between libpq and the backend, too.
Back-patch of a59c79564 and 50f03473e into all supported branches.
David Steele
Discussion: https://postgr.es/m/f4b7bc55-97ac-9e69-7398-335e212f7743@pgmasters.net
Perl can be convinced to execute user-defined code during compilation
of a plperl function (or at least a plperlu function). That's not
such a big problem as long as the activity is confined within the
Perl interpreter, and it's not clear we could do anything about that
anyway. However, if such code tries to use plperl's SPI functions,
we have a bigger problem. In the first place, those functions are
likely to crash because current_call_data->prodesc isn't set up yet.
In the second place, because it isn't set up, we lack critical info
such as whether the function is supposed to be read-only. And in
the third place, this path allows code execution during function
validation, which is strongly discouraged because of the potential
for security exploits. Hence, reject execution of the SPI functions
until compilation is finished.
While here, add check_spi_usage_allowed() calls to various functions
that hadn't gotten the memo about checking that. I think that perhaps
plperl_sv_to_literal may have been intentionally omitted on the grounds
that it was safe at the time; but if so, the addition of transforms
functionality changed that. The others are more recently added and
seem to be flat-out oversights.
Per report from Mark Murawski. Back-patch to all supported branches.
Discussion: https://postgr.es/m/9acdf918-7fff-4f40-f750-2ffa84f083d2@intellasoft.net
When opening a WAL file smaller than XLOG_BLCKSZ (e.g. 0 bytes long) while
determining the wal_segment_size, pg_waldump checked errno, despite errno not
being set by the short read. Resulting in a bogus error message.
Author: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Discussion: https://postgr.es/m/20220214.181847.775024684568733277.horikyota.ntt@gmail.com
Backpatch: 11-, the bug was introducedin fc49e24fa
When cleaning up temporary objects during process exit the cleanup could fail
with:
FATAL: cannot fetch toast data without an active snapshot
The bug is caused by RemoveTempRelationsCallback() not setting up a
snapshot. If an object with toasted catalog data needs to be cleaned up,
init_toast_snapshot() could fail with the above error.
Most of the time however the the problem is masked due to cached catalog
snapshots being returned by GetOldestSnapshot(). But dropping an object can
cause catalog invalidations to be emitted. If no further catalog accesses are
necessary between the invalidation processing and the next toast datum
deletion, the bug becomes visible.
It's easy to miss this bug because it typically happens after clients
disconnect and the FATAL error just ends up in the log.
Luckily temporary table cleanup at the next use of the same temporary schema
or during DISCARD ALL does not have the same problem.
Fix the bug by pushing a snapshot in RemoveTempRelationsCallback(). Also add
isolation tests for temporary object cleanup, including objects with toasted
catalog data.
A future HEAD only commit will add more assertions.
Reported-By: Miles Delahunty
Author: Andres Freund
Discussion: https://postgr.es/m/CAOFAq3BU5Mf2TTvu8D9n_ZOoFAeQswuzk7yziAb7xuw_qyw5gw@mail.gmail.com
Backpatch: 10-
Following migration of Windows buildfarm members running TAP tests to
use of ucrt64 perl for those tests, special processing for msys perl is
no longer necessary and so is removed.
Backpatch to release 10
Discussion: https://postgr.es/m/c65a8781-77ac-ea95-d185-6db291e1baeb@dunslane.net
Pre-v13 versions only support logical replication from plain tables,
while v13 and later also allow partitioned tables to be published.
If you tried to subscribe an older server to such a publication,
you got "table XXX not found on publisher", which is pretty
unhelpful/confusing. Arrange to deliver a more on-point error
message. As commit c314c147c did in v13, remove the relkind check
from the query WHERE clause altogether, so that "not there"
is distinguishable from "wrong relkind".
Per report from Radoslav Nedyalkov. Patch v10-v12.
Discussion: https://postgr.es/m/2952568.1644876730@sss.pgh.pa.us
Currently, during UPDATE, the unchanged replica identity key attributes
are not logged separately because they are getting logged as part of the
new tuple. But if they are stored externally then the untoasted values are
not getting logged as part of the new tuple and logical replication won't
be able to replicate such UPDATEs. So we need to log such attributes as
part of the old_key_tuple during UPDATE.
Reported-by: Haiying Tang
Author: Dilip Kumar and Amit Kapila
Reviewed-by: Alvaro Herrera, Haiying Tang, Andres Freund
Backpatch-through: 10
Discussion: https://postgr.es/m/OS0PR01MB611342D0A92D4F4BF26C0F47FB229@OS0PR01MB6113.jpnprd01.prod.outlook.com
createplan.c tries to save a runtime projection step by specifying
a scan plan node's output as being exactly the table's columns, or
index's columns in the case of an index-only scan, if there is not a
reason to do otherwise. This logic did not previously pay attention
to whether an index's columns are returnable. That worked, sort of
accidentally, until commit 9a3ddeb51 taught setrefs.c to reject plans
that try to read a non-returnable column. I have no desire to loosen
setrefs.c's new check, so instead adjust use_physical_tlist() to not
try to optimize this way when there are non-returnable column(s).
Per report from Ryan Kelly. Like the previous patch, back-patch
to all supported branches.
Discussion: https://postgr.es/m/CAHUie24ddN+pDNw7fkhNrjrwAX=fXXfGZZEHhRuofV_N_ftaSg@mail.gmail.com
"pg_ctl stop/restart" checked that the postmaster PID is valid just
once, as a side-effect of sending the stop signal, and then would
wait-till-timeout for the postmaster.pid file to go away. This
neglects the case wherein the postmaster dies uncleanly after we
signal it. Similarly, once "pg_ctl promote" has sent the signal,
it'd wait for the corresponding on-disk state change to occur
even if the postmaster dies.
I'm not sure how we've managed not to notice this problem, but it
seems to explain slow execution of the 017_shm.pl test script on AIX
since commit 4fdbf9af5, which added a speculative "pg_ctl stop" with
the idea of making real sure that the postmaster isn't there. In the
test steps that kill-9 and then restart the postmaster, it's possible
to get past the initial signal attempt before kill() stops working
for the doomed postmaster. If that happens, pg_ctl waited till
PGCTLTIMEOUT before giving up ... and the buildfarm's AIX members
have that set very high.
To fix, include a "kill(pid, 0)" test (similar to what
postmaster_is_alive uses) in these wait loops, so that we'll
give up immediately if the postmaster PID disappears.
While here, I chose to refactor those loops out of where they were.
do_stop() and do_restart() can perfectly well share one copy of the
wait-for-stop loop, and it seems desirable to put a similar function
beside that for wait-for-promote.
Back-patch to all supported versions, since pg_ctl's wait logic
is substantially identical in all, and we're seeing the slow test
behavior in all branches.
Discussion: https://postgr.es/m/20220210023537.GA3222837@rfd.leadboat.com
Some pre-2017 Test::More versions need perfect $Test::Builder::Level
maintenance to find the variable. Buildfarm member snapper reported an
overall failure that the file intended to hide via the TODO construct.
That trouble was reachable in v11 and v10. For later branches, this
serves as defense in depth. Back-patch to v10 (all supported versions).
Discussion: https://postgr.es/m/20220202055556.GB2745933@rfd.leadboat.com
The back-patch of commit fdd965d074 broke
CLOBBER_CACHE_ALWAYS for v9.6 through v13. It updated the
InvalidateSystemCaches() call for CLOBBER_CACHE_RECURSIVELY, neglecting
the one for CLOBBER_CACHE_ALWAYS. Back-patch to v13, v12, v11, and v10.
Reviewed by Tomas Vondra. Reported by Tomas Vondra.
Discussion: https://postgr.es/m/df7b4c0b-7d92-f03f-75c4-9e08b269a716@enterprisedb.com
Recent versions of Devel::PPPort try to redefine eval_pv() to
dodge a bug in pre-5.31 Perl versions. Unfortunately the redefinition
fails on compilers that don't support statements nested within
expressions. However, we aren't actually interested in this bug fix,
since we always call eval_pv() with croak_on_error = FALSE.
So, until there's an upstream fix for this breakage, just comment
out the macro to revert to the older behavior.
Per report from Wei Sun, as well as previous buildfarm failure
on pademelon (which I'd unfortunately not looked at carefully
enough to understand the cause). Back-patch to all supported
versions, since we're using the same ppport.h in all.
Discussion: https://postgr.es/m/tencent_2EFCC8BA0107B6EC0F97179E019A8A43C806@qq.com
Report: https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=pademelon&dt=2022-02-02%2001%3A22%3A58
A very well-informed user might deduce this from what we said already,
but I'd bet against it. Lay it out explicitly.
While here, rewrite the comment about tuple routing to be more
intelligible to an average SQL user.
Per bug #17395 from Alexander Lakhin. Back-patch to v11. (The text
in this area is different in v10 and I'm not sufficiently excited
about this point to adapt the patch.)
Discussion: https://postgr.es/m/17395-8c326292078d1a57@postgresql.org
There are two Asserts in nodeMergejoin.c that are reachable if
the input data is not in the expected order. This seems way too
fragile. Alexander Lakhin reported a case where the assertions
could be triggered with misconfigured foreign-table partitions,
and bitter experience with unstable operating system collation
definitions suggests another easy route to hitting them. Neither
Assert is in a place where we can't afford one more test-and-branch,
so replace 'em with plain test-and-elog logic.
Per bug #17395. While the reported symptom is relatively recent,
collation changes could happen anytime, so back-patch to all
supported branches.
Discussion: https://postgr.es/m/17395-8c326292078d1a57@postgresql.org
With Python 3.10, configure spits out warnings about the module
distutils.sysconfig being deprecated and scheduled for removal in
Python 3.12. Change the uses in configure to use the module sysconfig
instead. The logic stays largely the same, although we have to
rely on INCLUDEPY instead of the deprecated get_python_inc function.
Note that sysconfig exists since Python 2.7, so this moves the
minimum required version up from Python 2.6 (or 2.4, before v13).
Also, sysconfig didn't exist in Python 3.1, so the minimum 3.x
version is now 3.2.
Back-patch of commit bd233bdd8 into all supported branches.
In v10, this also includes back-patching v11's beff4bb9c, primarily
because this opinion is clearly out-of-date:
While at it, get rid of the code's assumption that both the major and
minor numbers contain exactly one digit. That will foreseeably be
broken by Python 3.10 in perhaps four or five years. That's far enough
out that we probably don't need to back-patch this.
Peter Eisentraut, Tom Lane, Andres Freund
Discussion: https://postgr.es/m/c74add3c-09c4-a9dd-1a03-a846e5b2fc52@enterprisedb.com
ppport.h was only updated in 05798c9f7f (master). Unfortunately my commit
c89f409749 uses PERL_VERSION_LT which came in with that update. Breaking most
buildfarm animals.
I should have noticed that...
We might want to backpatch the ppport update instead, but for now lets get the
buildfarm green again.
Discussion: https://postgr.es/m/20220131015130.shn6wr2fzuymerf6@alap3.anarazel.de
Backpatch: 10-14, master doesn't need it
For older versions we need our own copy of perl's setlocale(), because it was
not exposed (why we need the setlocale in the first place is explained in
plperl_init_interp) . The copy stopped working in 5.28, as some of the used
macros are not public anymore. But Perl_setlocale is available in 5.28, so
use that.
Author: Victor Wagner <vitus@wagner.pp.ru>
Reviewed-By: Dagfinn Ilmari Mannsåker <ilmari@ilmari.org>
Discussion: https://postgr.es/m/20200501134711.08750c5f@antares.wagner.home
Backpatch: all versions
Commit 8431e296ea reworked ProcArrayApplyRecoveryInfo to sort XIDs
before adding them to KnownAssignedXids. But the XIDs are sorted using
xidComparator, which compares the XIDs simply as uint32 values, not
logically. KnownAssignedXidsAdd() however expects XIDs in logical order,
and calls TransactionIdFollowsOrEquals() to enforce that. If there are
XIDs for which the two orderings disagree, an error is raised and the
recovery fails/restarts.
Hitting this issue is fairly easy - you just need two transactions, one
started before the 4B limit (e.g. XID 4294967290), the other sometime
after it (e.g. XID 1000). Logically (4294967290 <= 1000) but when
compared using xidComparator we try to add them in the opposite order.
Which makes KnownAssignedXidsAdd() fail with an error like this:
ERROR: out-of-order XID insertion in KnownAssignedXids
This only happens during replica startup, while processing RUNNING_XACTS
records to build the snapshot. Once we reach STANDBY_SNAPSHOT_READY, we
skip these records. So this does not affect already running replicas,
but if you restart (or create) a replica while there are transactions
with XIDs for which the two orderings disagree, you may hit this.
Long-running transactions and frequent replica restarts increase the
likelihood of hitting this issue. Once the replica gets into this state,
it can't be started (even if the old transactions are terminated).
Fixed by sorting the XIDs logically - this is fine because we're dealing
with normal XIDs (because it's XIDs assigned to backends) and from the
same wraparound epoch (otherwise the backends could not be running at
the same time on the primary node). So there are no problems with the
triangle inequality, which is why xidComparator compares raw values.
Investigation and root cause analysis by Abhijit Menon-Sen. Patch by me.
This issue is present in all releases since 9.4, however releases up to
9.6 are EOL already so backpatch to 10 only.
Reviewed-by: Abhijit Menon-Sen
Reviewed-by: Alvaro Herrera
Backpatch-through: 10
Discussion: https://postgr.es/m/36b8a501-5d73-277c-4972-f58a4dce088a%40enterprisedb.com
This reverts commits 6051857fc and ed52c3707, but only in the back
branches. Further testing has shown that while those changes do fix
some things, they also break others; in particular, it looks like
walreceivers fail to detect walsender-initiated connection close
reliably if the walsender shuts down this way. We'll keep trying to
improve matters in HEAD, but it now seems unwise to push these changes
into stable releases.
Discussion: https://postgr.es/m/CA+hUKG+OeoETZQ=Qw5Ub5h3tmwQhBmDA=nuNO3KG=zWfUypFAw@mail.gmail.com
Andres Freund
Alvaro Herrera
Michael Paquier
Tom Lane
Thomas Munro
Noah Misch
Andrew Dunstan
Etsuro Fujita
Amit Kapila
Alexander Korotkov
Peter Eisentraut
Bruce Momjian
Tomas Vondra