More precisely, what we do here is make the SLRU cache sizes
configurable with new GUCs, so that sites with high concurrency and big
ranges of transactions in flight (resp. multixacts/subtransactions) can
benefit from bigger caches. In order for this to work with good
performance, two additional changes are made:
1. the cache is divided in "banks" (to borrow terminology from CPU
caches), and algorithms such as eviction buffer search only affect
one specific bank. This forestalls the problem that linear searching
for a specific buffer across the whole cache takes too long: we only
have to search the specific bank, whose size is small. This work is
authored by Andrey Borodin.
2. Change the locking regime for the SLRU banks, so that each bank uses
a separate LWLock. This allows for increased scalability. This work
is authored by Dilip Kumar. (A part of this was previously committed as
d172b717c6f4.)
Special care is taken so that the algorithms that can potentially
traverse more than one bank release one bank's lock before acquiring the
next. This should happen rarely, but particularly clog.c's group commit
feature needed code adjustment to cope with this. I (Álvaro) also added
lots of comments to make sure the design is sound.
The new GUCs match the names introduced by bcdfa5f2e2 in the
pg_stat_slru view.
The default values for these parameters are similar to the previous
sizes of each SLRU. commit_ts, clog and subtrans accept value 0, which
means to adjust by dividing shared_buffers by 512 (so 2MB for every 1GB
of shared_buffers), with a cap of 8MB. (A new slru.c function
SimpleLruAutotuneBuffers() was added to support this.) The cap was
previously 1MB for clog, so for sites with more than 512MB of shared
memory the total memory used increases, which is likely a good tradeoff.
However, other SLRUs (notably multixact ones) retain smaller sizes and
don't support a configured value of 0. These values based on
shared_buffers may need to be revisited, but that's an easy change.
There was some resistance to adding these new GUCs: it would be better
to adjust to memory pressure automatically somehow, for example by
stealing memory from shared_buffers (where the caches can grow and
shrink naturally). However, doing that seems to be a much larger
project and one which has made virtually no progress in several years,
and because this is such a pain point for so many users, here we take
the pragmatic approach.
Author: Andrey Borodin <x4mmm@yandex-team.ru>
Author: Dilip Kumar <dilipbalaut@gmail.com>
Reviewed-by: Amul Sul, Gilles Darold, Anastasia Lubennikova,
Ivan Lazarev, Robert Haas, Thomas Munro, Tomas Vondra,
Yura Sokolov, Васильев Дмитрий (Dmitry Vasiliev).
Discussion: https://postgr.es/m/2BEC2B3F-9B61-4C1D-9FB5-5FAB0F05EF86@yandex-team.ru
Discussion: https://postgr.es/m/CAFiTN-vzDvNz=ExGXz6gdyjtzGixKSqs0mKHMmaQ8sOSEFZ33A@mail.gmail.com
There isn't a lot of user demand for AIX support, we have a bunch of
hacks to work around AIX-specific compiler bugs and idiosyncrasies,
and no one has stepped up to the plate to properly maintain it.
Remove support for AIX to get rid of that maintenance overhead. It's
still supported for stable versions.
The acute issue that triggered this decision was that after commit
8af2565248, the AIX buildfarm members have been hitting this
assertion:
TRAP: failed Assert("(uintptr_t) buffer == TYPEALIGN(PG_IO_ALIGN_SIZE, buffer)"), File: "md.c", Line: 472, PID: 2949728
Apperently the "pg_attribute_aligned(a)" attribute doesn't work on AIX
for values larger than PG_IO_ALIGN_SIZE, for a static const variable.
That could be worked around, but we decided to just drop the AIX support
instead.
Discussion: https://www.postgresql.org/message-id/20240224172345.32@rfd.leadboat.com
Reviewed-by: Andres Freund, Noah Misch, Thomas Munro
The new facility makes it easier to optimize bulk loading, as the
logic for buffering, WAL-logging, and syncing the relation only needs
to be implemented once. It's also less error-prone: We have had a
number of bugs in how a relation is fsync'd - or not - at the end of a
bulk loading operation. By centralizing that logic to one place, we
only need to write it correctly once.
The new facility is faster for small relations: Instead of of calling
smgrimmedsync(), we register the fsync to happen at next checkpoint,
which avoids the fsync latency. That can make a big difference if you
are e.g. restoring a schema-only dump with lots of relations.
It is also slightly more efficient with large relations, as the WAL
logging is performed multiple pages at a time. That avoids some WAL
header overhead. The sorted GiST index build did that already, this
moves the buffering to the new facility.
The changes to pageinspect GiST test needs an explanation: Before this
patch, the sorted GiST index build set the LSN on every page to the
special GistBuildLSN value, not the LSN of the WAL record, even though
they were WAL-logged. There was no particular need for it, it just
happened naturally when we wrote out the pages before WAL-logging
them. Now we WAL-log the pages first, like in B-tree build, so the
pages are stamped with the record's real LSN. When the build is not
WAL-logged, we still use GistBuildLSN. To make the test output
predictable, use an unlogged index.
Reviewed-by: Andres Freund
Discussion: https://www.postgresql.org/message-id/30e8f366-58b3-b239-c521-422122dd5150%40iki.fi
This commit adds timeout that is expected to be used as a prevention
of long-running queries. Any session within the transaction will be
terminated after spanning longer than this timeout.
However, this timeout is not applied to prepared transactions.
Only transactions with user connections are affected.
Discussion: https://postgr.es/m/CAAhFRxiQsRs2Eq5kCo9nXE3HTugsAAJdSQSmxncivebAxdmBjQ%40mail.gmail.com
Author: Andrey Borodin <amborodin@acm.org>
Author: Japin Li <japinli@hotmail.com>
Author: Junwang Zhao <zhjwpku@gmail.com>
Reviewed-by: Nikolay Samokhvalov <samokhvalov@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Reviewed-by: Fujii Masao <masao.fujii@oss.nttdata.com>
Reviewed-by: bt23nguyent <bt23nguyent@oss.nttdata.com>
Reviewed-by: Yuhang Qiu <iamqyh@gmail.com>
After calling smgropen(), it was not clear how long you could continue
to use the result, because various code paths including cache
invalidation could call smgrclose(), which freed the memory.
Guarantee that the object won't be destroyed until the end of the
current transaction, or in recovery, the commit/abort record that
destroys the underlying storage.
smgrclose() is now just an alias for smgrrelease(). It closes files
and forgets all state except the rlocator, but keeps the SMgrRelation
object valid.
A new smgrdestroy() function is used by rare places that know there
should be no other references to the SMgrRelation.
The short version:
* smgrclose() is now just an alias for smgrrelease(). It releases
resources, but doesn't destroy until EOX
* smgrdestroy() now frees memory, and should rarely be used.
Existing code should be unaffected, but it is now possible for code that
has an SMgrRelation object to use it repeatedly during a transaction as
long as the storage hasn't been physically dropped. Such code would
normally hold a lock on the relation.
This also replaces the "ownership" mechanism of SMgrRelations with a
pin counter. An SMgrRelation can now be "pinned", which prevents it
from being destroyed at end of transaction. There can be multiple pins
on the same SMgrRelation. In practice, the pin mechanism is only used
by the relcache, so there cannot be more than one pin on the same
SMgrRelation. Except with swap_relation_files XXX
Author: Thomas Munro, Heikki Linnakangas
Reviewed-by: Robert Haas <robertmhaas@gmail.com>
Discussion: https://www.postgresql.org/message-id/CA%2BhUKGJ8NTvqLHz6dqbQnt2c8XCki4r2QvXjBQcXpVwxTY_pvA@mail.gmail.com
Presently, the most straightforward way for a shared library to use
shared memory is to request it at server startup via a
shmem_request_hook, which requires specifying the library in
shared_preload_libraries. Alternatively, the library can create a
dynamic shared memory (DSM) segment, but absent a shared location
to store the segment's handle, other backends cannot use it. This
commit introduces a registry for DSM segments so that these other
backends can look up existing segments with a library-specified
string. This allows libraries to easily use shared memory without
needing to request it at server startup.
The registry is accessed via the new GetNamedDSMSegment() function.
This function handles allocating the segment and initializing it
via a provided callback. If another backend already created and
initialized the segment, it simply attaches the segment.
GetNamedDSMSegment() locks the registry appropriately to ensure
that only one backend initializes the segment and that all other
backends just attach it.
The registry itself is comprised of a dshash table that stores the
DSM segment handles keyed by a library-specified string.
Reviewed-by: Michael Paquier, Andrei Lepikhov, Nikita Malakhov, Robert Haas, Bharath Rupireddy, Zhang Mingli, Amul Sul
Discussion: https://postgr.es/m/20231205034647.GA2705267%40nathanxps13
jit.c and dfgr.c had a copy of the same code to check if a file exists
or not, with a twist: jit.c did not check for EACCES when failing the
stat() call for the path whose existence is tested. This refactored
routine will be used by an upcoming patch.
Reviewed-by: Ashutosh Bapat
Discussion: https://postgr.es/m/ZTiV8tn_MIb_H2rE@paquier.xyz
Both lwlocknames.txt and wait_event_names.txt contain a list of all
the predefined LWLocks, i.e., those with predefined positions
within MainLWLockArray. It is easy to miss one or the other,
especially since the list in wait_event_names.txt omits the "Lock"
suffix from all the LWLock wait events. This commit adds a cross-
check of these lists to the script that generates lwlocknames.h.
If the lists do not match exactly, building will fail.
Suggested-by: Robert Haas
Reviewed-by: Robert Haas, Michael Paquier, Bertrand Drouvot
Discussion: https://postgr.es/m/20240102173120.GA1061678%40nathanxps13
When active, this process writes WAL summary files to
$PGDATA/pg_wal/summaries. Each summary file contains information for a
certain range of LSNs on a certain TLI. For each relation, it stores a
"limit block" which is 0 if a relation is created or destroyed within
a certain range of WAL records, or otherwise the shortest length to
which the relation was truncated during that range of WAL records, or
otherwise InvalidBlockNumber. In addition, it stores a list of blocks
which have been modified during that range of WAL records, but
excluding blocks which were removed by truncation after they were
modified and never subsequently modified again.
In other words, it tells us which blocks need to copied in case of an
incremental backup covering that range of WAL records. But this
doesn't yet add the capability to actually perform an incremental
backup; the next patch will do that.
A new parameter summarize_wal enables or disables this new background
process. The background process also automatically deletes summary
files that are older than wal_summarize_keep_time, if that parameter
has a non-zero value and the summarizer is configured to run.
Patch by me, with some design help from Dilip Kumar and Andres Freund.
Reviewed by Matthias van de Meent, Dilip Kumar, Jakub Wartak, Peter
Eisentraut, and Álvaro Herrera.
Discussion: http://postgr.es/m/CA+TgmoYOYZfMCyOXFyC-P+-mdrZqm5pP2N7S-r0z3_402h9rsA@mail.gmail.com
smgrreadv() and smgrwritev() and their md.c implementations call
FileReadV() and FileWriteV(). A range of disk blocks beginning at
'blocknum' and extending for 'nblocks' can be scattered to or gathered
from multiple buffers with a single system call. The traditional
smgrread() and smgrwrite() functions are implemented in terms of the new
functions.
Later commits will introduce calls with nblocks > 1, but the following
behavioral changes can be seen already:
* After a short transfer we'll now retry until we eventually read 0
bytes (= EOF) or get ENOSPC, EDQUOT, EFBIG etc, where previously we
would infer the reason. Retrying is consistent with xlog.c's
treatment of large WAL writes, and arguably also xlog.c and fd.c's
treatment of EINTR. Arbitrary short returns for larger transfers have
been observed on several OSes, and might in theory also happen for
transient reasons with our own pg_p*v() fallback code.
* After unexpected EOF or -1, the error thrown now talks about
a range even for the single block case, eg "blocks 42..42".
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Discussion: https://postgr.es/m/CA+hUKGJkOiOCa+mag4BF+zHo7qo=o9CFheB8=g6uT5TUm2gkvA@mail.gmail.com
Previously smgrprefetch() could issue POSIX_FADV_WILLNEED advice for a
single block at a time. Add an nblocks argument so that we can do the
same for a range of blocks. This usually produces a single system call,
but might need to loop if it crosses a segment boundary. Initially it
is only called with nblocks == 1, but proposed patches will make wider
calls.
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi> (earlier version)
Discussion: https://postgr.es/m/CA+hUKGJkOiOCa+mag4BF+zHo7qo=o9CFheB8=g6uT5TUm2gkvA@mail.gmail.com
FileReadV() and FileWriteV() adapt pg_preadv() and pg_pwritev() for
fd.c's virtual file descriptors. The simple FileRead() and FileWrite()
functions are now implemented in terms of the vectored functions, to
avoid code duplication, and they are converted back to the corresponding
simple system calls further down (commit 15c9ac36). Later work will
make more interesting multi-iovec calls.
The traditional behavior of reporting a "fake" ENOSPC error is
simplified. It's now always set for non-failing writes, for the benefit
of callers that expect to log a meaningful "%m" if they determine that
the write was short. (Perhaps we should consider getting rid of that
expectation one day.)
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Discussion: https://postgr.es/m/CA+hUKGJkOiOCa+mag4BF+zHo7qo=o9CFheB8=g6uT5TUm2gkvA@mail.gmail.com
For clarity, have separate functions for *creating* the shared memory
and semaphores at postmaster or single-user backend startup, and
for *attaching* to existing shared memory structures in EXEC_BACKEND
case. CreateSharedMemoryAndSemaphores() is now called only at
postmaster startup, and a new AttachSharedMemoryStructs() function is
called at backend startup in EXEC_BACKEND mode.
Reviewed-by: Tristan Partin, Andres Freund
Discussion: https://www.postgresql.org/message-id/7a59b073-5b5b-151e-7ed3-8b01ff7ce9ef@iki.fi
We've had repeated bugs in the area of handling SLRU wraparound in the past,
some of which have caused data loss. Switching to an indexing system for SLRUs
that does not wrap around should allow us to get rid of a whole bunch
of problems and improve the overall reliability of the system.
This particular patch however only changes the indexing and doesn't address
the wraparound per se. This is going to be done in the following patches.
Author: Maxim Orlov, Aleksander Alekseev, Alexander Korotkov, Teodor Sigaev
Author: Nikita Glukhov, Pavel Borisov, Yura Sokolov
Reviewed-by: Jacob Champion, Heikki Linnakangas, Alexander Korotkov
Reviewed-by: Japin Li, Pavel Borisov, Tom Lane, Peter Eisentraut, Andres Freund
Reviewed-by: Andrey Borodin, Dilip Kumar, Aleksander Alekseev
Discussion: https://postgr.es/m/CACG%3DezZe1NQSCnfHOr78AtAZxJZeCvxrts0ygrxYwe%3DpyyjVWA%40mail.gmail.com
Discussion: https://postgr.es/m/CAJ7c6TPDOYBYrnCAeyndkBktO0WG2xSdYduTF0nxq%2BvfkmTF5Q%40mail.gmail.com
A WaitEventSet holds file descriptors or event handles (on Windows).
If FreeWaitEventSet is not called, those fds or handles are leaked.
Use ResourceOwners to track WaitEventSets, to clean those up
automatically on error.
This was a live bug in async Append nodes, if a FDW's
ForeignAsyncRequest function failed. (In back branches, I will apply a
more localized fix for that based on PG_TRY-PG_FINALLY.)
The added test doesn't check for leaking resources, so it passed even
before this commit. But at least it covers the code path.
In the passing, fix misleading comment on what the 'nevents' argument
to WaitEventSetWait means.
Report by Alexander Lakhin, analysis and suggestion for the fix by
Tom Lane. Fixes bug #17828.
Reviewed-by: Alexander Lakhin, Thomas Munro
Discussion: https://www.postgresql.org/message-id/472235.1678387869@sss.pgh.pa.us
The code previously relied on "long" as type to track block numbers,
which would be 4 bytes in all Windows builds or any 32-bit builds. This
limited the code to be able to handle up to 16TB of data with the
default block size of 8kB, like during a CLUSTER. This code now relies
on a more portable int64, which should be more than enough for at least
the next 20 years to come.
This issue has been reported back in 2017, but nothing was done about it
back then, so here we go now.
Reported-by: Peter Geoghegan
Reviewed-by: Heikki Linnakangas
Discussion: https://postgr.es/m/CAH2-WznCscXnWmnj=STC0aSa7QG+BRedDnZsP=Jo_R9GUZvUrg@mail.gmail.com
Instead of having a separate array/hash for each resource kind, use a
single array and hash to hold all kinds of resources. This makes it
possible to introduce new resource "kinds" without having to modify
the ResourceOwnerData struct. In particular, this makes it possible
for extensions to register custom resource kinds.
The old approach was to have a small array of resources of each kind,
and if it fills up, switch to a hash table. The new approach also uses
an array and a hash, but now the array and the hash are used at the
same time. The array is used to hold the recently added resources, and
when it fills up, they are moved to the hash. This keeps the access to
recent entries fast, even when there are a lot of long-held resources.
All the resource-specific ResourceOwnerEnlarge*(),
ResourceOwnerRemember*(), and ResourceOwnerForget*() functions have
been replaced with three generic functions that take resource kind as
argument. For convenience, we still define resource-specific wrapper
macros around the generic functions with the old names, but they are
now defined in the source files that use those resource kinds.
The release callback no longer needs to call ResourceOwnerForget on
the resource being released. ResourceOwnerRelease unregisters the
resource from the owner before calling the callback. That needed some
changes in bufmgr.c and some other files, where releasing the
resources previously always called ResourceOwnerForget.
Each resource kind specifies a release priority, and
ResourceOwnerReleaseAll releases the resources in priority order. To
make that possible, we have to restrict what you can do between
phases. After calling ResourceOwnerRelease(), you are no longer
allowed to remember any more resources in it or to forget any
previously remembered resources by calling ResourceOwnerForget. There
was one case where that was done previously. At subtransaction commit,
AtEOSubXact_Inval() would handle the invalidation messages and call
RelationFlushRelation(), which temporarily increased the reference
count on the relation being flushed. We now switch to the parent
subtransaction's resource owner before calling AtEOSubXact_Inval(), so
that there is a valid ResourceOwner to temporarily hold that relcache
reference.
Other end-of-xact routines make similar calls to AtEOXact_Inval()
between release phases, but I didn't see any regression test failures
from those, so I'm not sure if they could reach a codepath that needs
remembering extra resources.
There were two exceptions to how the resource leak WARNINGs on commit
were printed previously: llvmjit silently released the context without
printing the warning, and a leaked buffer io triggered a PANIC. Now
everything prints a WARNING, including those cases.
Add tests in src/test/modules/test_resowner.
Reviewed-by: Aleksander Alekseev, Michael Paquier, Julien Rouhaud
Reviewed-by: Kyotaro Horiguchi, Hayato Kuroda, Álvaro Herrera, Zhihong Yu
Reviewed-by: Peter Eisentraut, Andres Freund
Discussion: https://www.postgresql.org/message-id/cbfabeb0-cd3c-e951-a572-19b365ed314d%40iki.fi
Since C99, there can be a trailing comma after the last value in an
enum definition. A lot of new code has been introducing this style on
the fly. Some new patches are now taking an inconsistent approach to
this. Some add the last comma on the fly if they add a new last
value, some are trying to preserve the existing style in each place,
some are even dropping the last comma if there was one. We could
nudge this all in a consistent direction if we just add the trailing
commas everywhere once.
I omitted a few places where there was a fixed "last" value that will
always stay last. I also skipped the header files of libpq and ecpg,
in case people want to use those with older compilers. There were
also a small number of cases where the enum type wasn't used anywhere
(but the enum values were), which ended up confusing pgindent a bit,
so I left those alone.
Discussion: https://www.postgresql.org/message-id/flat/386f8c45-c8ac-4681-8add-e3b0852c1620%40eisentraut.org
Enforce the rule from transam/README in XLogRegisterBuffer(), and
update callers to follow the rule.
Hash indexes sometimes register clean pages as a part of the locking
protocol, so provide a REGBUF_NO_CHANGE flag to support that use.
Discussion: https://postgr.es/m/c84114f8-c7f1-5b57-f85a-3adc31e1a904@iki.fi
Reviewed-by: Heikki Linnakangas
Instead of returning the number of characters in the RelFileNumber,
return the RelFileNumber itself. Continue to return the fork number,
as before, and additionally return the segment number.
parse_filename_for_nontemp_relation now rejects a RelFileNumber or
segment number that begins with a leading zero. Before, we accepted
such cases as relation filenames, but if we continued to do so after
this change, the function might return the same values for two
different files (e.g. 1234.5 and 001234.5 or 1234.005) which could be
annoying for callers. Since we don't actually ever generate filenames
with leading zeroes in the names, any such files that we find must
have been created by something other than PostgreSQL, and it is
therefore reasonable to treat them as non-relation files.
Along the way, change unlogged_relation_entry to store a RelFileNumber
rather than an OID. This update should have been made in
851f4cc75c, but it was overlooked.
It's trivial to make the update as part of this commit, perhaps more
trivial than it would have been without it, so do that.
Patch by me, reviewed by David Steele.
Discussion: http://postgr.es/m/CA+TgmoZNVeBzoqDL8xvr-nkaepq815jtDR4nJzPew7=3iEuM1g@mail.gmail.com
This commit introduces trigger on login event, allowing to fire some actions
right on the user connection. This can be useful for logging or connection
check purposes as well as for some personalization of environment. Usage
details are described in the documentation included, but shortly usage is
the same as for other triggers: create function returning event_trigger and
then create event trigger on login event.
In order to prevent the connection time overhead when there are no triggers
the commit introduces pg_database.dathasloginevt flag, which indicates database
has active login triggers. This flag is set by CREATE/ALTER EVENT TRIGGER
command, and unset at connection time when no active triggers found.
Author: Konstantin Knizhnik, Mikhail Gribkov
Discussion: https://postgr.es/m/0d46d29f-4558-3af9-9c85-7774e14a7709%40postgrespro.ru
Reviewed-by: Pavel Stehule, Takayuki Tsunakawa, Greg Nancarrow, Ivan Panchenko
Reviewed-by: Daniel Gustafsson, Teodor Sigaev, Robert Haas, Andres Freund
Reviewed-by: Tom Lane, Andrey Sokolov, Zhihong Yu, Sergey Shinderuk
Reviewed-by: Gregory Stark, Nikita Malakhov, Ted Yu
We shouldn't be doing non-trivial work in signal handlers in general,
and in this case the handler could reach unsafe code and corrupt state.
It also clobbered its own "reason" code.
Move all recovery conflict decision logic into the next
CHECK_FOR_INTERRUPTS(), and have the signal handler just set flags and
the latch, following the standard pattern. Since there are several
different "reasons", use a separate flag for each.
With this refactoring, the recovery conflict system no longer
piggy-backs on top of the regular query cancelation mechanism, but
instead raises an error directly if it decides that is necessary. It
still needs to respect QueryCancelHoldoffCount, because otherwise the
FEBE protocol might get out of sync (see commit 2b3a8b20c2).
This fixes one class of intermittent failure in the new
031_recovery_conflict.pl test added by commit 9f8a050f, though the buggy
coding is much older. Failures outside contrived testing seem to be
very rare (or perhaps incorrectly attributed) in the field, based on
lack of reports.
No back-patch for now due to complexity and release schedule. We have
the option to back-patch into 16 later, as 16 has prerequisite commit
bea3d7e.
Reviewed-by: Andres Freund <andres@anarazel.de> (earlier version)
Reviewed-by: Michael Paquier <michael@paquier.xyz> (earlier version)
Reviewed-by: Robert Haas <robertmhaas@gmail.com> (earlier version)
Tested-by: Christoph Berg <myon@debian.org>
Discussion: https://postgr.es/m/CA%2BhUKGK3PGKwcKqzoosamn36YW-fsuTdOPPF1i_rtEO%3DnEYKSg%40mail.gmail.com
Discussion: https://postgr.es/m/CALj2ACVr8au2J_9D88UfRCi0JdWhyQDDxAcSVav0B0irx9nXEg%40mail.gmail.com
This commit renames RecoveryInitSyncMethod to DataDirSyncMethod and
moves it to common/file_utils.h. This is preparatory work for a
follow-up commit that will allow specifying the synchronization
method in frontend utilities such as pg_upgrade and pg_basebackup.
Reviewed-by: Michael Paquier
Discussion: https://postgr.es/m/ZN2ZB4afQ2JbR9TA%40paquier.xyz
Presently, frontend code that needs to use these macros must either
include storage/fd.h, which declares several frontend-unsafe
functions, or duplicate the macros. This commit moves these macros
to common/file_utils.h, which is safe for both frontend and backend
code. Consequently, we can also remove the duplicated macros in
pg_checksums and stop including storage/fd.h in pg_rewind.
Reviewed-by: Michael Paquier
Discussion: https://postgr.es/m/ZOP5qoUualu5xl2Z%40paquier.xyz
Commit 31966b15 invented a way for functions dealing with relation
extension to accept a Relation in online code and an SMgrRelation in
recovery code. It seems highly likely that future bufmgr.c interfaces
will face the same problem, and need to do something similar.
Generalize the names so that each interface doesn't have to re-invent
the wheel.
Back-patch to 16. Since extension AM authors might start using the
constructor macros once 16 ships, we agreed to do the rename in 16
rather than waiting for 17.
Reviewed-by: Peter Geoghegan <pg@bowt.ie>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CA%2BhUKG%2B6tLD2BhpRWycEoti6LVLyQq457UL4ticP5xd8LqHySA%40mail.gmail.com
Commit b91dd9de was concerned with a theoretical problem with our
non-atomic condition variable operations. If you stop sleeping, and
then cancel the sleep in a separate step, you might be signaled in
between, and that could be lost. That doesn't matter for callers of
ConditionVariableBroadcast(), but callers of ConditionVariableSignal()
might be upset if a signal went missing like this.
Commit bc971f4025 interacted badly with that logic, because it doesn't
use ConditionVariableSleep(), which would normally put us back in the
wait list. ConditionVariableCancelSleep() would be confused and think
we'd received an extra signal, and try to forward it to another backend,
resulting in wakeup storms.
New idea: ConditionVariableCancelSleep() can just return true if we've
been signaled. Hypothetical users of ConditionVariableSignal() would
then still have a way to deal with rare lost signals if they are
concerned about that problem.
Back-patch to 16, where bc971f4025 arrived.
Reported-by: Tomas Vondra <tomas.vondra@enterprisedb.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/2840876b-4cfe-240f-0a7e-29ffd66711e7%40enterprisedb.com
The WAL insertion lock variable insertingAt is currently being read
and written with the help of the LWLock wait list lock to avoid any read
of torn values. This wait list lock can become a point of contention on
a highly concurrent write workloads.
This commit switches insertingAt to a 64b atomic variable that provides
torn-free reads/writes. On platforms without 64b atomic support, the
fallback implementation uses spinlocks to provide the same guarantees
for the values read. LWLockWaitForVar(), through
LWLockConflictsWithVar(), reads the new value to check if it still needs
to wait with a u64 atomic operation. LWLockUpdateVar() updates the
variable before waking up the waiters with an exchange_u64 (full memory
barrier). LWLockReleaseClearVar() now uses also an exchange_u64 to
reset the variable. Before this commit, all these steps relied on
LWLockWaitListLock() and LWLockWaitListUnlock().
This reduces contention on LWLock wait list lock and improves
performance of highly-concurrent write workloads. Here are some
numbers using pg_logical_emit_message() (HEAD at d6677b93) with various
arbitrary record lengths and clients up to 1k on a rather-large machine
(64 vCPUs, 512GB of RAM, 16 cores per sockets, 2 sockets), in terms of
TPS numbers coming from pgbench:
message_size_b | 16 | 64 | 256 | 1024
--------------------+--------+--------+--------+-------
patch_4_clients | 83830 | 82929 | 80478 | 73131
patch_16_clients | 267655 | 264973 | 250566 | 213985
patch_64_clients | 380423 | 378318 | 356907 | 294248
patch_256_clients | 360915 | 354436 | 326209 | 263664
patch_512_clients | 332654 | 321199 | 287521 | 240128
patch_1024_clients | 288263 | 276614 | 258220 | 217063
patch_2048_clients | 252280 | 243558 | 230062 | 192429
patch_4096_clients | 212566 | 213654 | 205951 | 166955
head_4_clients | 83686 | 83766 | 81233 | 73749
head_16_clients | 266503 | 265546 | 249261 | 213645
head_64_clients | 366122 | 363462 | 341078 | 261707
head_256_clients | 132600 | 132573 | 134392 | 165799
head_512_clients | 118937 | 114332 | 116860 | 150672
head_1024_clients | 133546 | 115256 | 125236 | 151390
head_2048_clients | 137877 | 117802 | 120909 | 138165
head_4096_clients | 113440 | 115611 | 120635 | 114361
Bharath has been measuring similar improvements, where the limit of the
WAL insertion lock begins to be felt when more than 256 concurrent
clients are involved in this specific workload.
An extra patch has been discussed to introduce a fast-exit path in
LWLockUpdateVar() when there are no waiters, still this does not
influence the write-heavy workload cases discussed as there are always
waiters. This will be considered separately.
Author: Bharath Rupireddy
Reviewed-by: Nathan Bossart, Andres Freund, Michael Paquier
Discussion: https://postgr.es/m/CALj2ACVF+6jLvqKe6xhDzCCkr=rfd6upaGc3477Pji1Ke9G7Bg@mail.gmail.com
This is useful to show the allocation state of huge pages when setting
up a server with "huge_pages = try", where allocating huge pages would
be attempted but the server would continue its startup sequence even if
the allocation fails. The effective status of huge pages is not easily
visible without OS-level tools (or for instance, a lookup at
/proc/N/smaps), and the environments where Postgres runs may not
authorize that. Like the other GUCs related to huge pages, this works
for Linux and Windows.
This GUC can report as values:
- "on", if huge pages were allocated.
- "off", if huge pages were not allocated.
- "unknown", a special state that could only be seen when using for
example postgres -C because it is only possible to know if the shared
memory allocation worked after we can check for the GUC values, even if
checking a runtime-computed GUC. This value should never be seen when
querying for the GUC on a running server. An assertion is added to
check that.
The discussion has also turned around having a new function to grab this
status, but this would have required more tricks for -DEXEC_BACKEND,
something that GUCs already handle.
Noriyoshi Shinoda has initiated the thread that has led to the result of
this commit.
Author: Justin Pryzby
Reviewed-by: Nathan Bossart, Kyotaro Horiguchi, Michael Paquier
Discussion: https://postgr.es/m/TU4PR8401MB1152EBB0D271F827E2E37A01EECC9@TU4PR8401MB1152.NAMPRD84.PROD.OUTLOOK.COM
Run pgindent, pgperltidy, and reformat-dat-files.
This set of diffs is a bit larger than typical. We've updated to
pg_bsd_indent 2.1.2, which properly indents variable declarations that
have multi-line initialization expressions (the continuation lines are
now indented one tab stop). We've also updated to perltidy version
20230309 and changed some of its settings, which reduces its desire to
add whitespace to lines to make assignments etc. line up. Going
forward, that should make for fewer random-seeming changes to existing
code.
Discussion: https://postgr.es/m/20230428092545.qfb3y5wcu4cm75ur@alvherre.pgsql
28e626bde0 added the concept of IOOps but neglected to include writeback
operations. ac8d53dae5 added time spent doing these I/O operations. Without
counting writeback, checkpointer write time in the log often differed
substantially from that in pg_stat_io. To fix this, add IOOp IOOP_WRITEBACK
and track writeback in pg_stat_io.
Bumps catversion.
Author: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Reported-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/20230419172326.dhgyo4wrrhulovt6%40awork3.anarazel.de
For clarity of review, renaming the function parameter "context" in
ScheduleBufferTagForWriteback() and IssuePendingWritebacks() to
"wb_context" is a separate commit. The next commit adds an "io_context"
parameter and "wb_context" makes it more clear which is which.
Author: Melanie Plageman <melanieplageman@gmail.com>
Discussion: https://postgr.es/m/CAAKRu_acc6iL4M3hvOTeztf_ZPpsB3Pqio5aVHgZ5q=Pi3BZKg@mail.gmail.com
The recently added inclusion of guc.h in smgr.h is not necessary and
introduces more server-related stuff. Removing the directive helps
avoid potential issues with including sgmr.h in frontends.
Author: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Discussion: https://postgr.es/m/20230425.115748.2130383825066921512.horikyota.ntt%40gmail.com
vacuum_defer_cleanup_age was introduced before hot_standby_feedback and
replication slots existed. It is hard to use reasonably - commonly it will
either be set too low (not preventing recovery conflicts, while still causing
some bloat), or too high (causing a lot of bloat). The alternatives do not
have that issue.
That on its own might not be sufficient reason to remove
vacuum_defer_cleanup_age, but it also complicates computation of xid
horizons. See e.g. the bug fixed in be504a3e97. It also is untested.
This commit removes TransactionIdRetreatSafely(), as there are no users
anymore. There might be potential future users, hence noting that here.
Reviewed-by: Daniel Gustafsson <daniel@yesql.se>
Reviewed-by: Justin Pryzby <pryzby@telsasoft.com>
Reviewed-by: Alvaro Herrera <alvherre@alvh.no-ip.org>
Discussion: https://postgr.es/m/20230317230930.nhsgk3qfk7f4axls@awork3.anarazel.de
Make sure that function declarations use names that exactly match the
corresponding names from function definitions in a few places. These
inconsistencies were all introduced relatively recently, after the code
base had parameter name mismatches fixed in bulk (see commits starting
with commits 4274dc22 and 035ce1fe).
pg_bsd_indent still has a couple of similar inconsistencies, which I
(pgeoghegan) have left untouched for now.
Like all earlier commits that cleaned up function parameter names, this
commit was written with help from clang-tidy.
During WAL replay on the standby, when a conflict with a logical slot is
identified, invalidate such slots. There are two sources of conflicts:
1) Using the information added in 6af1793954, logical slots are invalidated if
required rows are removed
2) wal_level on the primary server is reduced to below logical
Uses the infrastructure introduced in the prior commit. FIXME: add commit
reference.
Change InvalidatePossiblyObsoleteSlot() to use a recovery conflict to
interrupt use of a slot, if called in the startup process. The new recovery
conflict is added to pg_stat_database_conflicts, as confl_active_logicalslot.
See 6af1793954 for an overall design of logical decoding on a standby.
Bumps catversion for the addition of the pg_stat_database_conflicts column.
Bumps PGSTAT_FILE_FORMAT_ID for the same reason.
Author: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>
Author: Andres Freund <andres@anarazel.de>
Author: Amit Khandekar <amitdkhan.pg@gmail.com> (in an older version)
Reviewed-by: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Reviewed-by: Robert Haas <robertmhaas@gmail.com>
Reviewed-by: Fabrízio de Royes Mello <fabriziomello@gmail.com>
Reviewed-by: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Reviewed-by: Amit Kapila <amit.kapila16@gmail.com>
Reviewed-by: Alvaro Herrera <alvherre@alvh.no-ip.org>
Discussion: https://postgr.es/m/20230407075009.igg7be27ha2htkbt@awork3.anarazel.de
Provide a way to ask the kernel to use O_DIRECT (or local equivalent)
where available for data and WAL files, to avoid or minimize kernel
caching. This hurts performance currently and is not intended for end
users yet. Later proposed work would introduce our own I/O clustering,
read-ahead, etc to replace the facilities the kernel disables with this
option.
The only user-visible change, if the developer-only GUC is not used, is
that this commit also removes the obscure logic that would activate
O_DIRECT for the WAL when wal_sync_method=open_[data]sync and
wal_level=minimal (which also requires max_wal_senders=0). Those are
non-default and unlikely settings, and this behavior wasn't (correctly)
documented. The same effect can be achieved with io_direct=wal.
Author: Thomas Munro <thomas.munro@gmail.com>
Author: Andres Freund <andres@anarazel.de>
Author: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Reviewed-by: Justin Pryzby <pryzby@telsasoft.com>
Reviewed-by: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Discussion: https://postgr.es/m/CA%2BhUKGK1X532hYqJ_MzFWt0n1zt8trz980D79WbjwnT-yYLZpg%40mail.gmail.com
In order to have the option to use O_DIRECT/FILE_FLAG_NO_BUFFERING in a
later commit, we need the addresses of user space buffers to be well
aligned. The exact requirements vary by OS and file system (typically
sectors and/or memory pages). The address alignment size is set to
4096, which is enough for currently known systems: it matches modern
sectors and common memory page size. There is no standard governing
O_DIRECT's requirements so we might eventually have to reconsider this
with more information from the field or future systems.
Aligning I/O buffers on memory pages is also known to improve regular
buffered I/O performance.
Three classes of I/O buffers for regular data pages are adjusted:
(1) Heap buffers are now allocated with the new palloc_aligned() or
MemoryContextAllocAligned() functions introduced by commit 439f6175.
(2) Stack buffers now use a new struct PGIOAlignedBlock to respect
PG_IO_ALIGN_SIZE, if possible with this compiler. (3) The buffer
pool is also aligned in shared memory.
WAL buffers were already aligned on XLOG_BLCKSZ. It's possible for
XLOG_BLCKSZ to be configured smaller than PG_IO_ALIGNED_SIZE and thus
for O_DIRECT WAL writes to fail to be well aligned, but that's a
pre-existing condition and will be addressed by a later commit.
BufFiles are not yet addressed (there's no current plan to use O_DIRECT
for those, but they could potentially get some incidental speedup even
in plain buffered I/O operations through better alignment).
If we can't align stack objects suitably using the compiler extensions
we know about, we disable the use of O_DIRECT by setting PG_O_DIRECT to
0. This avoids the need to consider systems that have O_DIRECT but
can't align stack objects the way we want; such systems could in theory
be supported with more work but we don't currently know of any such
machines, so it's easier to pretend there is no O_DIRECT support
instead. That's an existing and tested class of system.
Add assertions that all buffers passed into smgrread(), smgrwrite() and
smgrextend() are correctly aligned, unless PG_O_DIRECT is 0 (= stack
alignment tricks may be unavailable) or the block size has been set too
small to allow arrays of buffers to be all aligned.
Author: Thomas Munro <thomas.munro@gmail.com>
Author: Andres Freund <andres@anarazel.de>
Reviewed-by: Justin Pryzby <pryzby@telsasoft.com>
Discussion: https://postgr.es/m/CA+hUKGK1X532hYqJ_MzFWt0n1zt8trz980D79WbjwnT-yYLZpg@mail.gmail.com
Add new options to the VACUUM and ANALYZE commands called
BUFFER_USAGE_LIMIT to allow users more control over how large to make the
buffer access strategy that is used to limit the usage of buffers in
shared buffers. Larger rings can allow VACUUM to run more quickly but
have the drawback of VACUUM possibly evicting more buffers from shared
buffers that might be useful for other queries running on the database.
Here we also add a new GUC named vacuum_buffer_usage_limit which controls
how large to make the access strategy when it's not specified in the
VACUUM/ANALYZE command. This defaults to 256KB, which is the same size as
the access strategy was prior to this change. This setting also
controls how large to make the buffer access strategy for autovacuum.
Per idea by Andres Freund.
Author: Melanie Plageman
Reviewed-by: David Rowley
Reviewed-by: Andres Freund
Reviewed-by: Justin Pryzby
Reviewed-by: Bharath Rupireddy
Discussion: https://postgr.es/m/20230111182720.ejifsclfwymw2reb@awork3.anarazel.de
The primary bottlenecks for relation extension are:
1) The extension lock is held while acquiring a victim buffer for the new
page. Acquiring a victim buffer can require writing out the old page
contents including possibly needing to flush WAL.
2) When extending via ReadBuffer() et al, we write a zero page during the
extension, and then later write out the actual page contents. This can
nearly double the write rate.
3) The existing bulk relation extension infrastructure in hio.c just amortized
the cost of acquiring the relation extension lock, but none of the other
costs.
Unfortunately 1) cannot currently be addressed in a central manner as the
callers to ReadBuffer() need to acquire the extension lock. To address that,
this this commit moves the responsibility for acquiring the extension lock
into bufmgr.c functions. That allows to acquire the relation extension lock
for just the required time. This will also allow us to improve relation
extension further, without changing callers.
The reason we write all-zeroes pages during relation extension is that we hope
to get ENOSPC errors earlier that way (largely works, except for CoW
filesystems). It is easier to handle out-of-space errors gracefully if the
page doesn't yet contain actual tuples. This commit addresses 2), by using the
recently introduced smgrzeroextend(), which extends the relation, without
dirtying the kernel page cache for all the extended pages.
To address 3), this commit introduces a function to extend a relation by
multiple blocks at a time.
There are three new exposed functions: ExtendBufferedRel() for extending the
relation by a single block, ExtendBufferedRelBy() to extend a relation by
multiple blocks at once, and ExtendBufferedRelTo() for extending a relation up
to a certain size.
To avoid duplicating code between ReadBuffer(P_NEW) and the new functions,
ReadBuffer(P_NEW) now implements relation extension with
ExtendBufferedRel(), using a flag to tell ExtendBufferedRel() that the
relation lock is already held.
Note that this commit does not yet lead to a meaningful performance or
scalability improvement - for that uses of ReadBuffer(P_NEW) will need to be
converted to ExtendBuffered*(), which will be done in subsequent commits.
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-by: Melanie Plageman <melanieplageman@gmail.com>
Discussion: https://postgr.es/m/20221029025420.eplyow6k7tgu6he3@awork3.anarazel.de
A future patch will add support for extending relations by multiple blocks at
once. To be concurrency safe, the buffers for those blocks need to be marked
as BM_IO_IN_PROGRESS. Until now we only had infrastructure for recovering from
an IO error for a single buffer. This commit extends that infrastructure to
multiple buffers by using the resource owner infrastructure.
This commit increases the size of the ResourceOwnerData struct, which appears
to have a just about measurable overhead in very extreme workloads. Medium
term we are planning to substantially shrink the size of
ResourceOwnerData. Short term the increase is small enough to not worry about
it for now.
Reviewed-by: Melanie Plageman <melanieplageman@gmail.com>
Discussion: https://postgr.es/m/20221029025420.eplyow6k7tgu6he3@awork3.anarazel.de
Discussion: https://postgr.es/m/20221029200025.w7bvlgvamjfo6z44@awork3.anarazel.de
This adds a few more assertions against a buffer being local in places we
don't expect, and extracts the check for a buffer being pinned exactly once
from LockBufferForCleanup() into its own function. Later commits will use this
function.
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-by: Melanie Plageman <melanieplageman@gmail.com>
Discussion: http://postgr.es/m/419312fd-9255-078c-c3e3-f0525f911d7f@iki.fi
smgrzeroextend() uses FileFallocate() to efficiently extend files by multiple
blocks. When extending by a small number of blocks, use FileZero() instead, as
using posix_fallocate() for small numbers of blocks is inefficient for some
file systems / operating systems. FileZero() is also used as the fallback for
FileFallocate() on platforms / filesystems that don't support fallocate.
A big advantage of using posix_fallocate() is that it typically won't cause
dirty buffers in the kernel pagecache. So far the most common pattern in our
code is that we smgrextend() a page full of zeroes and put the corresponding
page into shared buffers, from where we later write out the actual contents of
the page. If the kernel, e.g. due to memory pressure or elapsed time, already
wrote back the all-zeroes page, this can lead to doubling the amount of writes
reaching storage.
There are no users of smgrzeroextend() as of this commit. That will follow in
future commits.
Reviewed-by: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Reviewed-by: David Rowley <dgrowleyml@gmail.com>
Reviewed-by: John Naylor <john.naylor@enterprisedb.com>
Discussion: https://postgr.es/m/20221029025420.eplyow6k7tgu6he3@awork3.anarazel.de
Among other things, this should make it easier to calculate a useful cache hit
ratio by excluding buffer reads via buffer access strategies. As buffer access
strategies reuse buffers (and thus evict the prior buffer contents), it is
normal to see reads on repeated scans of the same data.
Author: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CAAKRu_beMa9Hzih40%3DXPYqhDVz6tsgUGTrhZXRo%3Dunp%2Bszb%3DUA%40mail.gmail.com
This function has been a no-op for over a decade. Even if bufmgr
regains a need to be called during commit, it seems unlikely that
the most appropriate call points would be precisely here, so it's not
doing us much good as a placeholder either. Now, removing it probably
doesn't save any noticeable number of cycles --- but the main call is
inside the commit critical section, and the less work done there the
better.
Matthias van de Meent
Discussion: https://postgr.es/m/CAEze2Wi1=tLKbxZnXzcD+8fYKyKqBtivVakLQC_mYBsP4Y8qVA@mail.gmail.com
Commit 28e626bde0 introduced the infrastructure for tracking more detailed IO
statistics. This commit adds the actual collection of the new IO statistics
for relations and temporary relations. See aforementioned commit for goals and
high-level design.
The changes in this commit are fairly straight-forward. The bulk of the change
is to passing sufficient information to the callsites of pgstat_count_io_op().
A somewhat unsightly detail is that it currently is hard to find a better
place to count fsyncs than in md.c, whereas the other pgstat_count_io_op()
calls are in bufmgr.c/localbuf.c. As the number of fsyncs is tied to md.c
implementation details, it's not obvious there is a better answer.
Author: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/20200124195226.lth52iydq2n2uilq@alap3.anarazel.de
Invent separate macros for "invalid" values of these types, so that
we needn't embed knowledge of their representations into calling code.
These are all zeroes anyway ATM, so this is not fixing any live bug,
but it makes the code cleaner and more future-proof.
I (tgl) also chose to move DSM_HANDLE_INVALID into dsm_impl.h,
since it seems like it should live beside the typedef for dsm_handle.
Hou Zhijie, Nathan Bossart, Kyotaro Horiguchi, Tom Lane
Discussion: https://postgr.es/m/OS0PR01MB5716860B1454C34E5B179B6694C99@OS0PR01MB5716.jpnprd01.prod.outlook.com
Enforce wal_retrieve_retry_interval on a per-subscription basis,
rather than globally, and arrange to skip that delay in case of
an intentional worker exit. This probably makes little difference
in the field, where apply workers wouldn't be restarted often;
but it has a significant impact on the runtime of our logical
replication regression tests (even though those tests use
artificially-small wal_retrieve_retry_interval settings already).
Nathan Bossart, with mostly-cosmetic editorialization by me
Discussion: https://postgr.es/m/20221122004119.GA132961@nathanxps13
This provides a way to reserve connection slots for non-superusers.
The slots reserved via the new GUC are available only to users who
have the new predefined role pg_use_reserved_connections.
superuser_reserved_connections remains as a final reserve in case
reserved_connections has been exhausted.
Patch by Nathan Bossart. Reviewed by Tushar Ahuja and by me.
Discussion: http://postgr.es/m/20230119194601.GA4105788@nathanxps13
This makes sure that the internal logic of these functions does not
attempt to change the value of the arguments constified, and it removes
one unconstify() in basic_archive.c.
Author: Nathan Bossart
Reviewed-by: Andrew Dunstan, Peter Eisentraut
Discussion: https://postgr.es/m/20230114231126.GA2580330@nathanxps13
Most callers of BufFileRead() want to check whether they read the full
specified length. Checking this at every call site is very tedious.
This patch provides additional variants BufFileReadExact() and
BufFileReadMaybeEOF() that include the length checks.
I considered changing BufFileRead() itself, but this function is also
used in extensions, and so changing the behavior like this would
create a lot of problems there. The new names are analogous to the
existing LogicalTapeReadExact().
Reviewed-by: Amit Kapila <amit.kapila16@gmail.com>
Discussion: https://www.postgresql.org/message-id/flat/f3501945-c591-8cc3-5ef0-b72a2e0eaa9c@enterprisedb.com
Currently, for large transactions, the publisher sends the data in
multiple streams (changes divided into chunks depending upon
logical_decoding_work_mem), and then on the subscriber-side, the apply
worker writes the changes into temporary files and once it receives the
commit, it reads from those files and applies the entire transaction. To
improve the performance of such transactions, we can instead allow them to
be applied via parallel workers.
In this approach, we assign a new parallel apply worker (if available) as
soon as the xact's first stream is received and the leader apply worker
will send changes to this new worker via shared memory. The parallel apply
worker will directly apply the change instead of writing it to temporary
files. However, if the leader apply worker times out while attempting to
send a message to the parallel apply worker, it will switch to
"partial serialize" mode - in this mode, the leader serializes all
remaining changes to a file and notifies the parallel apply workers to
read and apply them at the end of the transaction. We use a non-blocking
way to send the messages from the leader apply worker to the parallel
apply to avoid deadlocks. We keep this parallel apply assigned till the
transaction commit is received and also wait for the worker to finish at
commit. This preserves commit ordering and avoid writing to and reading
from files in most cases. We still need to spill if there is no worker
available.
This patch also extends the SUBSCRIPTION 'streaming' parameter so that the
user can control whether to apply the streaming transaction in a parallel
apply worker or spill the change to disk. The user can set the streaming
parameter to 'on/off', or 'parallel'. The parameter value 'parallel' means
the streaming will be applied via a parallel apply worker, if available.
The parameter value 'on' means the streaming transaction will be spilled
to disk. The default value is 'off' (same as current behaviour).
In addition, the patch extends the logical replication STREAM_ABORT
message so that abort_lsn and abort_time can also be sent which can be
used to update the replication origin in parallel apply worker when the
streaming transaction is aborted. Because this message extension is needed
to support parallel streaming, parallel streaming is not supported for
publications on servers < PG16.
Author: Hou Zhijie, Wang wei, Amit Kapila with design inputs from Sawada Masahiko
Reviewed-by: Sawada Masahiko, Peter Smith, Dilip Kumar, Shi yu, Kuroda Hayato, Shveta Mallik
Discussion: https://postgr.es/m/CAA4eK1+wyN6zpaHUkCLorEWNx75MG0xhMwcFhvjqm2KURZEAGw@mail.gmail.com
An epoll fd belonging to the parent should be closed in the child. A
kqueue fd is automatically closed by fork(), but we should still adjust
our counter. For poll and Windows systems, nothing special is required.
On all systems we free the memory.
No caller yet, but we'll need this if we start using WaitEventSet in the
postmaster as planned.
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CA%2BhUKG%2BZ-HpOj1JsO9eWUP%2Bar7npSVinsC_npxSy%2BjdOMsx%3DGg%40mail.gmail.com
To be able to handle incoming connections on a server socket with
the WaitEventSet API, we'll need a new kind of event to indicate that
the the socket is ready to accept a connection.
On Unix, it's just the same as WL_SOCKET_READABLE, but on Windows there
is a different underlying kernel event that we need to map our
abstraction to.
No user yet, but a proposed patch would use this.
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CA%2BhUKG%2BZ-HpOj1JsO9eWUP%2Bar7npSVinsC_npxSy%2BjdOMsx%3DGg%40mail.gmail.com
A new function pg_stat_get_backend_subxact() can be used to get
information about the number of subtransactions in the cache of
a particular backend and whether that cache has overflowed. This
can be useful for tracking down performance problems that can
result from overflowed snapshots.
Dilip Kumar, reviewed by Zhihong Yu, Nikolay Samokhvalov,
Justin Pryzby, Nathan Bossart, Ashutosh Sharma, Julien
Rouhaud. Additional design comments from Andres Freund,
Tom Lane, Bruce Momjian, and David G. Johnston.
Discussion: http://postgr.es/m/CAFiTN-ut0uwkRJDQJeDPXpVyTWD46m3gt3JDToE02hTfONEN=Q@mail.gmail.com
Because we added StaticAssertStmt() first before StaticAssertDecl(),
some uses as well as the instructions in c.h are now a bit backwards
from the "native" way static assertions are meant to be used in C.
This updates the guidance and moves some static assertions to better
places.
Specifically, since the addition of StaticAssertDecl(), we can put
static assertions at the file level. This moves a number of static
assertions out of function bodies, where they might have been stuck
out of necessity, to perhaps better places at the file level or in
header files.
Also, when the static assertion appears in a position where a
declaration is allowed, then using StaticAssertDecl() is more native
than StaticAssertStmt().
Reviewed-by: John Naylor <john.naylor@enterprisedb.com>
Discussion: https://www.postgresql.org/message-id/flat/941a04e7-dd6f-c0e4-8cdf-a33b3338cbda%40enterprisedb.com
Make the argument types of the File API match stdio better:
- Change the data buffer to void *, from char *.
- Change FileWrite() data buffer to const on top of that.
- Change amounts to size_t, from int.
In passing, change the FilePrefetch() amount argument from int to
off_t, to match the underlying posix_fadvise().
Discussion: https://www.postgresql.org/message-id/flat/11dda853-bb5b-59ba-a746-e168b1ce4bdb%40enterprisedb.com
Previously, we'd compress only when the active range of array entries
reached Max(4 * PROCARRAY_MAXPROCS, 2 * pArray->numKnownAssignedXids).
If max_connections is large, the first term could result in not
compressing for a long time, resulting in much wastage of cycles in
hot-standby backends scanning the array to take snapshots. Get rid
of that term, and just bound it to 2 * pArray->numKnownAssignedXids.
That however creates the opposite risk, that we might spend too much
effort compressing. Hence, consider compressing only once every 128
commit records. (This frequency was chosen by benchmarking. While
we only tried one benchmark scenario, the results seem stable over
a fairly wide range of frequencies.)
Also, force compression when processing RecoveryInfo WAL records
(which should be infrequent); the old code could perform compression
then, but would do so only after the same array-range check as for
the transaction-commit path.
Also, opportunistically run compression if the startup process is about
to wait for WAL, though not oftener than once a second. This should
prevent cases where we waste lots of time by leaving the array
not-compressed for long intervals due to low WAL traffic.
Lastly, add a simple check to keep us from uselessly compressing
when the array storage is already compact.
Back-patch, as the performance problem is worse in pre-v14 branches
than in HEAD.
Simon Riggs and Michail Nikolaev, with help from Tom Lane and
Andres Freund.
Discussion: https://postgr.es/m/CALdSSPgahNUD_=pB_j=1zSnDBaiOtqVfzo8Ejt5J_k7qZiU1Tw@mail.gmail.com
Until now LWLockDequeueSelf() sequentially searched the list of waiters to see
if the current proc is still is on the list of waiters, or has already been
removed. In extreme workloads, where the wait lists are very long, this leads
to a quadratic behavior. #backends iterating over a list #backends
long. Additionally, the likelihood of needing to call LWLockDequeueSelf() in
the first place also increases with the increased length of the wait queue, as
it becomes more likely that a lock is released while waiting for the wait list
lock, which is held for longer during lock release.
Due to the exponential back-off in perform_spin_delay() this is surprisingly
hard to detect. We should make that easier, e.g. by adding a wait event around
the pg_usleep() - but that's a separate patch.
The fix is simple - track whether a proc is currently waiting in the wait list
or already removed but waiting to be woken up in PGPROC->lwWaiting.
In some workloads with a lot of clients contending for a small number of
lwlocks (e.g. WALWriteLock), the fix can substantially increase throughput.
As the quadratic behavior arguably is a bug, we might want to decide to
backpatch this fix in the future.
Author: Andres Freund <andres@anarazel.de>
Reviewed-by: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Discussion: https://postgr.es/m/20221027165914.2hofzp4cvutj6gin@awork3.anarazel.de
Discussion: https://postgr.es/m/CALj2ACXktNbG=K8Xi7PSqbofTZozavhaxjatVc14iYaLu4Maag@mail.gmail.com
Standardize on the name snapshotConflictHorizon for all XID fields from
WAL records that generate recovery conflicts when in hot standby mode.
This supersedes the previous latestRemovedXid naming convention.
The new naming convention places emphasis on how the values are actually
used by REDO routines. How the values are generated during original
execution (details of which vary by record type) is deemphasized. Users
of tools like pg_waldump can now grep for snapshotConflictHorizon to see
all potential sources of recovery conflicts in a standardized way,
without necessarily having to consider which specific record types might
be involved.
Also bring a couple of WAL record types that didn't follow any kind of
naming convention into line. These are heapam's VISIBLE record type and
SP-GiST's VACUUM_REDIRECT record type. Now every WAL record whose REDO
routine calls ResolveRecoveryConflictWithSnapshot() passes through the
snapshotConflictHorizon field from its WAL record. This is follow-up
work to the refactoring from commit 9e540599 that made FREEZE_PAGE WAL
records use a standard snapshotConflictHorizon style XID cutoff.
No bump in XLOG_PAGE_MAGIC, since the underlying format of affected WAL
records doesn't change.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CAH2-Wzm2CQUmViUq7Opgk=McVREHSOorYaAjR1ZpLYkRN7_dPw@mail.gmail.com
If we have no special-case code in s_lock.h for the current platform,
but the compiler has __sync_lock_test_and_set, use that instead of
failing. It's unlikely that anybody's __sync_lock_test_and_set
would be so awful as to be worse than our semaphore-based fallback,
but if it is, they can (continue to) use --disable-spinlocks.
This allows removal of the RISC-V special case installed by commit
c32fcac56, which generated exactly the same code but only on that
platform. Usefully, the RISC-V buildfarm animals should now test
at least the int variant of this patch.
I've manually tested both variants on ARM by dint of removing the
ARM-specific stanza. We don't want to drop that, because it already
has some special knowledge and is likely to grow more over time.
Likewise, this is not meant to preclude installing special cases
for other arches if that proves worthwhile.
Per discussion of a request to install the same code for loongarch64.
Like the previous patch, we might as well back-patch to supported
branches.
Discussion: https://postgr.es/m/761ac43d44b84d679ba803c2bd947cc0@HSMAILSVR04.hs.handsome.com.cn
This is similar to 7d25958, and this commit takes care of all the
remaining inconsistencies between the initial value used in the C
variable associated to a GUC and its default value stored in the GUC
tables (as of pg_settings.boot_val).
Some of the initial values of the GUCs updated rely on a compile-time
default. These are refactored so as the GUC table and its C declaration
use the same values. This makes everything consistent with other
places, backend_flush_after, bgwriter_flush_after, port,
checkpoint_flush_after doing so already, for example.
Extracted from a larger patch by Peter Smith. The spots updated in the
modules are from me.
Author: Peter Smith, Michael Paquier
Reviewed-by: Nathan Bossart, Tom Lane, Justin Pryzby
Discussion: https://postgr.es/m/CAHut+PtHE0XSfjjRQ6D4v7+dqzCw=d+1a64ujra4EX8aoc_Z+w@mail.gmail.com
This commit moves pg_pwritev_with_retry(), a convenience wrapper of
pg_writev() able to handle partial writes, to common/file_utils.c so
that the frontend code is able to use it. A first use-case targetted
for this routine is pg_basebackup and pg_receivewal, for the
zero-padding of a newly-initialized WAL segment. This is used currently
in the backend when the GUC wal_init_zero is enabled (default).
Author: Bharath Rupireddy
Reviewed-by: Nathan Bossart, Thomas Munro
Discussion: https://postgr.es/m/CALj2ACUq7nAb7=bJNbK3yYmp-SZhJcXFR_pLk8un6XgDzDF3OA@mail.gmail.com
There are still some alignment-related failures in the buildfarm,
which might or might not be able to be fixed quickly, but I've also
just realized that it increased the size of many WAL records by 4 bytes
because a block reference contains a RelFileLocator. The effect of that
hasn't been studied or discussed, so revert for now.
SharedInvalSmgrMsg can't require 8-byte alignment, because then
SharedInvalidationMessage will require 8-byte alignment, which will
then cause ParseCommitRecord to fail on machines that are picky
about alignment, because it assumes that everything that gets
packed into a commit record requires only 4-byte alignment.
Another problem with 05d4cbf9b6.
Discussion: http://postgr.es/m/3825454.1664310917@sss.pgh.pa.us
RelFileNumbers are now assigned using a separate counter, instead of
being assigned from the OID counter. This counter never wraps around:
if all 2^56 possible RelFileNumbers are used, an internal error
occurs. As the cluster is limited to 2^64 total bytes of WAL, this
limitation should not cause a problem in practice.
If the counter were 64 bits wide rather than 56 bits wide, we would
need to increase the width of the BufferTag, which might adversely
impact buffer lookup performance. Also, this lets us use bigint for
pg_class.relfilenode and other places where these values are exposed
at the SQL level without worrying about overflow.
This should remove the need to keep "tombstone" files around until
the next checkpoint when relations are removed. We do that to keep
RelFileNumbers from being recycled, but now that won't happen
anyway. However, this patch doesn't actually change anything in
this area; it just makes it possible for a future patch to do so.
Dilip Kumar, based on an idea from Andres Freund, who also reviewed
some earlier versions of the patch. Further review and some
wordsmithing by me. Also reviewed at various points by Ashutosh
Sharma, Vignesh C, Amul Sul, Álvaro Herrera, and Tom Lane.
Discussion: http://postgr.es/m/CA+Tgmobp7+7kmi4gkq7Y+4AM9fTvL+O1oQ4-5gFTT+6Ng-dQ=g@mail.gmail.com
Autoconf is showing its age, fewer and fewer contributors know how to wrangle
it. Recursive make has a lot of hard to resolve dependency issues and slow
incremental rebuilds. Our home-grown MSVC build system is hard to maintain for
developers not using Windows and runs tests serially. While these and other
issues could individually be addressed with incremental improvements, together
they seem best addressed by moving to a more modern build system.
After evaluating different build system choices, we chose to use meson, to a
good degree based on the adoption by other open source projects.
We decided that it's more realistic to commit a relatively early version of
the new build system and mature it in tree.
This commit adds an initial version of a meson based build system. It supports
building postgres on at least AIX, FreeBSD, Linux, macOS, NetBSD, OpenBSD,
Solaris and Windows (however only gcc is supported on aix, solaris). For
Windows/MSVC postgres can now be built with ninja (faster, particularly for
incremental builds) and msbuild (supporting the visual studio GUI, but
building slower).
Several aspects (e.g. Windows rc file generation, PGXS compatibility, LLVM
bitcode generation, documentation adjustments) are done in subsequent commits
requiring further review. Other aspects (e.g. not installing test-only
extensions) are not yet addressed.
When building on Windows with msbuild, builds are slower when using a visual
studio version older than 2019, because those versions do not support
MultiToolTask, required by meson for intra-target parallelism.
The plan is to remove the MSVC specific build system in src/tools/msvc soon
after reaching feature parity. However, we're not planning to remove the
autoconf/make build system in the near future. Likely we're going to keep at
least the parts required for PGXS to keep working around until all supported
versions build with meson.
Some initial help for postgres developers is at
https://wiki.postgresql.org/wiki/Meson
With contributions from Thomas Munro, John Naylor, Stone Tickle and others.
Author: Andres Freund <andres@anarazel.de>
Author: Nazir Bilal Yavuz <byavuz81@gmail.com>
Author: Peter Eisentraut <peter@eisentraut.org>
Reviewed-By: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Discussion: https://postgr.es/m/20211012083721.hvixq4pnh2pixr3j@alap3.anarazel.de
Make sure that function declarations use names that exactly match the
corresponding names from function definitions in storage, catalog,
access method, executor, and logical replication code, as well as in
miscellaneous utility/library code.
Like other recent commits that cleaned up function parameter names, this
commit was written with help from clang-tidy. Later commits will do the
same for other parts of the codebase.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/CAH2-WznJt9CMM9KJTMjJh_zbL5hD9oX44qdJ4aqZtjFi-zA3Tg@mail.gmail.com
This is preparatory work for a project to increase the number of bits
in a RelFileNumber from 32 to 56.
Along the way, introduce static inline accessor functions for a couple
of BufferTag fields.
Dilip Kumar, reviewed by me. The overall patch series has also had
review at various times from Andres Freund, Ashutosh Sharma, Hannu
Krosing, Vignesh C, Álvaro Herrera, and Tom Lane.
Discussion: http://postgr.es/m/CAFiTN-trubju5YbWAq-BSpZ90-Z6xCVBQE8BVqXqANOZAF1Znw@mail.gmail.com
Assume that we can use LWARX hint flags and the LWSYNC instruction
on any PPC machine. The check on the assembler's behavior was only
needed for Apple's old assembler, which is no longer of interest
now that we've de-supported all PPC-era versions of macOS (thanks
to them not having clock_gettime()). Also, given an up-to-date
assembler these instructions work even on Apple's old hardware.
It seems quite unlikely that anyone would be interested in running
current Postgres on PPC hardware that's so old as to not have
these instructions.
Hence, rip out associated configure test and manual configuration
options, and just use the modernized instructions all the time.
Also, update atomics/arch-ppc.h to use these instructions as well.
(It was already using LWSYNC unconditionally in another place,
providing further proof that nobody is using PG on hardware old
enough to have a problem with that.)
Discussion: https://postgr.es/m/166622.1660323391@sss.pgh.pa.us
Previously we fell back to __FUNCTION__ and then NULL. As __func__ is in C99
that shouldn't be necessary anymore.
Solution.pm defined HAVE_FUNCNAME__FUNCTION instead of
HAVE_FUNCNAME__FUNC (originating in 4164e6636e), as at some point in the past
MSVC only supported __FUNCTION__. Our minimum version supports __func__.
Reviewed-By: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/20220807012914.ydz73yte6j3coulo@awork3.anarazel.de
Commit b0a55e4329 missed a few places
where we are referring to the number used as a part of the relation
filename as an "OID". We now want to call that a "RelFileNumber".
Some of these places actually made it sound like the OID in question
is pg_class.oid rather than pg_class.relfilenode, which is especially
good to clean up.
Dilip Kumar with some editing by me.
