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
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
We have been using the term RelFileNode to refer to either (1) the
integer that is used to name the sequence of files for a certain relation
within the directory set aside for that tablespace/database combination;
or (2) that value plus the OIDs of the tablespace and database; or
occasionally (3) the whole series of files created for a relation
based on those values. Using the same name for more than one thing is
confusing.
Replace RelFileNode with RelFileNumber when we're talking about just the
single number, i.e. (1) from above, and with RelFileLocator when we're
talking about all the things that are needed to locate a relation's files
on disk, i.e. (2) from above. In the places where we refer to (3) as
a relfilenode, instead refer to "relation storage".
Since there is a ton of SQL code in the world that knows about
pg_class.relfilenode, don't change the name of that column, or of other
SQL-facing things that derive their name from it.
On the other hand, do adjust closely-related internal terminology. For
example, the structure member names dbNode and spcNode appear to be
derived from the fact that the structure itself was called RelFileNode,
so change those to dbOid and spcOid. Likewise, various variables with
names like rnode and relnode get renamed appropriately, according to
how they're being used in context.
Hopefully, this is clearer than before. It is also preparation for
future patches that intend to widen the relfilenumber fields from its
current width of 32 bits. Variables that store a relfilenumber are now
declared as type RelFileNumber rather than type Oid; right now, these
are the same, but that can now more easily be changed.
Dilip Kumar, per an idea from me. Reviewed also by Andres Freund.
I fixed some whitespace issues, changed a couple of words in a
comment, and made one other minor correction.
Discussion: http://postgr.es/m/CA+TgmoamOtXbVAQf9hWFzonUo6bhhjS6toZQd7HZ-pmojtAmag@mail.gmail.com
Discussion: http://postgr.es/m/CA+Tgmobp7+7kmi4gkq7Y+4AM9fTvL+O1oQ4-5gFTT+6Ng-dQ=g@mail.gmail.com
Discussion: http://postgr.es/m/CAFiTN-vTe79M8uDH1yprOU64MNFE+R3ODRuA+JWf27JbhY4hJw@mail.gmail.com
Up until now, we've had a policy of only marking certain variables
in the PostgreSQL header files with PGDLLIMPORT, but now we've
decided to mark them all. This means that extensions running on
Windows should no longer operate at a disadvantage as compared to
extensions running on Linux: if the variable is present in a header
file, it should be accessible.
Discussion: http://postgr.es/m/CA+TgmoYanc1_FSfimhgiWSqVyP5KKmh5NP2BWNwDhO8Pg2vGYQ@mail.gmail.com
Instead of referring to target backends by pid, use pgprocno. This
means that we don't have to scan the ProcArray and we can drop some
special case code for dealing with the startup process.
Discussion: https://postgr.es/m/CA%2BhUKGLYRyDaneEwz5Uya_OgFLMx5BgJfkQSD%3Dq9HmwsfRRb-w%40mail.gmail.com
Reviewed-by: Soumyadeep Chakraborty <soumyadeep2007@gmail.com>
Reviewed-by: Ashwin Agrawal <ashwinstar@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
1. Backends waiting for buffer I/O are now interruptible.
2. If something goes wrong in a backend that is currently performing
I/O, waiting backends no longer wake up until that backend reaches
AbortBufferIO() and broadcasts on the CV. Previously, any waiters would
wake up (because the I/O lock was automatically released) and then
busy-loop until AbortBufferIO() cleared BM_IO_IN_PROGRESS.
3. LWLockMinimallyPadded is removed, as it would now be unused.
Author: Robert Haas <robertmhaas@gmail.com>
Reviewed-by: Thomas Munro <thomas.munro@gmail.com>
Reviewed-by: Julien Rouhaud <rjuju123@gmail.com>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us> (earlier version, 2016)
Discussion: https://postgr.es/m/CA%2BhUKGJ8nBFrjLuCTuqKN0pd2PQOwj9b_jnsiGFFMDvUxahj_A%40mail.gmail.com
Discussion: https://postgr.es/m/CA+Tgmoaj2aPti0yho7FeEf2qt-JgQPRWb0gci_o1Hfr=C56Xng@mail.gmail.com
Provide PrefetchSharedBuffer(), a variant that takes SMgrRelation, for
use in recovery. Rename LocalPrefetchBuffer() to PrefetchLocalBuffer()
for consistency.
Add a return value to all of these. In recovery, tolerate and report
missing files, so we can handle relations unlinked before crash recovery
began. Also report cache hits and misses, so that callers can do faster
buffer lookups and better I/O accounting.
Reviewed-by: Alvaro Herrera <alvherre@2ndquadrant.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CA%2BhUKGJ4VJN8ttxScUFM8dOKX0BrBiboo5uz1cq%3DAovOddfHpA%40mail.gmail.com
Periodically while the server is running, and at shutdown, write out a
list of blocks in shared buffers. When the server reaches consistency
-- unfortunatey, we can't do it before that point without breaking
things -- reload those blocks into any still-unused shared buffers.
Mithun Cy and Robert Haas, reviewed and tested by Beena Emerson,
Amit Kapila, Jim Nasby, and Rafia Sabih.
Discussion: http://postgr.es/m/CAD__OugubOs1Vy7kgF6xTjmEqTR4CrGAv8w+ZbaY_+MZeitukw@mail.gmail.com
Change pg_bsd_indent to follow upstream rules for placement of comments
to the right of code, and remove pgindent hack that caused comments
following #endif to not obey the general rule.
Commit e3860ffa4d wasn't actually using
the published version of pg_bsd_indent, but a hacked-up version that
tried to minimize the amount of movement of comments to the right of
code. The situation of interest is where such a comment has to be
moved to the right of its default placement at column 33 because there's
code there. BSD indent has always moved right in units of tab stops
in such cases --- but in the previous incarnation, indent was working
in 8-space tab stops, while now it knows we use 4-space tabs. So the
net result is that in about half the cases, such comments are placed
one tab stop left of before. This is better all around: it leaves
more room on the line for comment text, and it means that in such
cases the comment uniformly starts at the next 4-space tab stop after
the code, rather than sometimes one and sometimes two tabs after.
Also, ensure that comments following #endif are indented the same
as comments following other preprocessor commands such as #else.
That inconsistency turns out to have been self-inflicted damage
from a poorly-thought-through post-indent "fixup" in pgindent.
This patch is much less interesting than the first round of indent
changes, but also bulkier, so I thought it best to separate the effects.
Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org
Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
This could result in corruption of the init fork of an unlogged index
if the ambuildempty routine for that index used shared buffers to
create the init fork, which was true for brin, gin, gist, and hash
indexes.
Patch by me, based on an earlier patch by Michael Paquier, who also
reviewed this one. This also incorporates an idea from Artur
Zakirov.
Discussion: http://postgr.es/m/CACYUyc8yccE4xfxhqxfh_Mh38j7dRFuxfaK1p6dSNAEUakxUyQ@mail.gmail.com
I somehow had assumed that in the spinlock (in turn possibly using
semaphores) based fallback atomics implementation 32 bit writes could be
done without a lock. As far as the write goes that's correct, since
postgres supports only platforms with single-copy atomicity for aligned
32bit writes. But writing without holding the spinlock breaks
read-modify-write operations like pg_atomic_compare_exchange_u32(),
since they'll potentially "miss" a concurrent write, which can't happen
in actual hardware implementations.
In 9.6+ when using the fallback atomics implementation this could lead
to buffer header locks not being properly marked as released, and
potentially some related state corruption. I don't see a related danger
in 9.5 (earliest release with the API), because pg_atomic_write_u32()
wasn't used in a concurrent manner there.
The state variable of local buffers, before this change, were
manipulated using pg_atomic_write_u32(), to avoid unnecessary
synchronization overhead. As that'd not be the case anymore, introduce
and use pg_atomic_unlocked_write_u32(), which does not correctly
interact with RMW operations.
This bug only caused issues when postgres is compiled on platforms
without atomics support (i.e. no common new platform), or when compiled
with --disable-atomics, which explains why this wasn't noticed in
testing.
Reported-By: Tom Lane
Discussion: <14947.1475690465@sss.pgh.pa.us>
Backpatch: 9.5-, where the atomic operations API was introduced.
This bug appears to have been introduced late in the development of
48354581a4 ("Allow Pin/UnpinBuffer to operate in a lockfree
manner.").
Found while debugging a bug which turned out to be independent of the
commit mentioned above.
Backpatch: -
The recent patch to make Pin/UnpinBuffer lockfree in the hot
path (48354581a), accidentally used pg_atomic_fetch_or_u32() in
MarkLocalBufferDirty(). Other code operating on local buffers was
careful to only use pg_atomic_read/write_u32 which just read/write from
memory; to avoid unnecessary overhead.
On its own that'd just make MarkLocalBufferDirty() slightly less
efficient, but in addition InitLocalBuffers() doesn't call
pg_atomic_init_u32() - thus the spinlock fallback for the atomic
operations isn't initialized. That in turn caused, as reported by Tom,
buildfarm animal gaur to fail. As those errors are actually useful
against this type of error, continue to omit - intentionally this time -
initialization of the atomic variable.
In addition, add an explicit note about only using pg_atomic_read/write
on local buffers's state to BufferDesc's description.
Reported-By: Tom Lane
Discussion: 1881.1460431476@sss.pgh.pa.us
Pinning/Unpinning a buffer is a very frequent operation; especially in
read-mostly cache resident workloads. Benchmarking shows that in various
scenarios the spinlock protecting a buffer header's state becomes a
significant bottleneck. The problem can be reproduced with pgbench -S on
larger machines, but can be considerably worse for queries which touch
the same buffers over and over at a high frequency (e.g. nested loops
over a small inner table).
To allow atomic operations to be used, cram BufferDesc's flags,
usage_count, buf_hdr_lock, refcount into a single 32bit atomic variable;
that allows to manipulate them together using 32bit compare-and-swap
operations. This requires reducing MAX_BACKENDS to 2^18-1 (which could
be lifted by using a 64bit field, but it's not a realistic configuration
atm).
As not all operations can easily implemented in a lockfree manner,
implement the previous buf_hdr_lock via a flag bit in the atomic
variable. That way we can continue to lock the header in places where
it's needed, but can get away without acquiring it in the more frequent
hot-paths. There's some additional operations which can be done without
the lock, but aren't in this patch; but the most important places are
covered.
As bufmgr.c now essentially re-implements spinlocks, abstract the delay
logic from s_lock.c into something more generic. It now has already two
users, and more are coming up; there's a follupw patch for lwlock.c at
least.
This patch is based on a proof-of-concept written by me, which Alexander
Korotkov made into a fully working patch; the committed version is again
revised by me. Benchmarking and testing has, amongst others, been
provided by Dilip Kumar, Alexander Korotkov, Robert Haas.
On a large x86 system improvements for readonly pgbench, with a high
client count, of a factor of 8 have been observed.
Author: Alexander Korotkov and Andres Freund
Discussion: 2400449.GjM57CE0Yg@dinodell
Up to now checkpoints were written in the order they're in the
BufferDescriptors. That's nearly random in a lot of cases, which
performs badly on rotating media, but even on SSDs it causes slowdowns.
To avoid that, sort checkpoints before writing them out. We currently
sort by tablespace, relfilenode, fork and block number.
One of the major reasons that previously wasn't done, was fear of
imbalance between tablespaces. To address that balance writes between
tablespaces.
The other prime concern was that the relatively large allocation to sort
the buffers in might fail, preventing checkpoints from happening. Thus
pre-allocate the required memory in shared memory, at server startup.
This particularly makes it more efficient to have checkpoint flushing
enabled, because that'll often result in a lot of writes that can be
coalesced into one flush.
Discussion: alpine.DEB.2.10.1506011320000.28433@sto
Author: Fabien Coelho and Andres Freund
Currently writes to the main data files of postgres all go through the
OS page cache. This means that some operating systems can end up
collecting a large number of dirty buffers in their respective page
caches. When these dirty buffers are flushed to storage rapidly, be it
because of fsync(), timeouts, or dirty ratios, latency for other reads
and writes can increase massively. This is the primary reason for
regular massive stalls observed in real world scenarios and artificial
benchmarks; on rotating disks stalls on the order of hundreds of seconds
have been observed.
On linux it is possible to control this by reducing the global dirty
limits significantly, reducing the above problem. But global
configuration is rather problematic because it'll affect other
applications; also PostgreSQL itself doesn't always generally want this
behavior, e.g. for temporary files it's undesirable.
Several operating systems allow some control over the kernel page
cache. Linux has sync_file_range(2), several posix systems have msync(2)
and posix_fadvise(2). sync_file_range(2) is preferable because it
requires no special setup, whereas msync() requires the to-be-flushed
range to be mmap'ed. For the purpose of flushing dirty data
posix_fadvise(2) is the worst alternative, as flushing dirty data is
just a side-effect of POSIX_FADV_DONTNEED, which also removes the pages
from the page cache. Thus the feature is enabled by default only on
linux, but can be enabled on all systems that have any of the above
APIs.
While desirable and likely possible this patch does not contain an
implementation for windows.
With the infrastructure added, writes made via checkpointer, bgwriter
and normal user backends can be flushed after a configurable number of
writes. Each of these sources of writes controlled by a separate GUC,
checkpointer_flush_after, bgwriter_flush_after and backend_flush_after
respectively; they're separate because the number of flushes that are
good are separate, and because the performance considerations of
controlled flushing for each of these are different.
A later patch will add checkpoint sorting - after that flushes from the
ckeckpoint will almost always be desirable. Bgwriter flushes are most of
the time going to be random, which are slow on lots of storage hardware.
Flushing in backends works well if the storage and bgwriter can keep up,
but if not it can have negative consequences. This patch is likely to
have negative performance consequences without checkpoint sorting, but
unfortunately so has sorting without flush control.
Discussion: alpine.DEB.2.10.1506011320000.28433@sto
Author: Fabien Coelho and Andres Freund
Move the content lock directly into the BufferDesc, so that locking and
pinning a buffer touches only one cache line rather than two. Adjust
the definition of BufferDesc slightly so that this doesn't make the
BufferDesc any larger than one cache line (at least on platforms where
a spinlock is only 1 or 2 bytes).
We can't fit the I/O locks into the BufferDesc and stay within one
cache line, so move those to a completely separate tranche. This
leaves a relatively limited number of LWLocks in the main tranche, so
increase the padding of those remaining locks to a full cache line,
rather than allowing adjacent locks to share a cache line, hopefully
reducing false sharing.
Performance testing shows that these changes make little difference
on laptop-class machines, but help significantly on larger servers,
especially those with more than 2 sockets.
Andres Freund, originally based on an earlier patch by Simon Riggs.
Review and cosmetic adjustments (including heavy rewriting of the
comments) by me.
Prior to commit 0709b7ee72, access to
variables within a spinlock-protected critical section had to be done
through a volatile pointer, but that should no longer be necessary.
Review by Andres Freund
Benchmarks has shown that aligning the buffer descriptor array to
cache lines is important for scalability; especially on bigger,
multi-socket, machines.
Currently the array sometimes already happens to be aligned by
happenstance, depending how large previous shared memory allocations
were. That can lead to wildly varying performance results after minor
configuration changes.
In addition to aligning the start of descriptor array, also force the
size of individual descriptors to be of a common cache line size (64
bytes). That happens to already be the case on 64bit platforms, but
this way we can change the struct BufferDesc more easily.
As the alignment primarily matters in highly concurrent workloads
which probably all are 64bit these days, and the space wastage of
element alignment would be a bit more noticeable on 32bit systems, we
don't force the stride to be cacheline sized on 32bit platforms for
now. If somebody does actual performance testing, we can reevaluate
that decision by changing the definition of BUFFERDESC_PADDED_SIZE.
Discussion: 20140202151319.GD32123@awork2.anarazel.de
Per discussion with Bruce Momjan, Tom Lane, Robert Haas, and Peter
Geoghegan.
StrategyGetBuffer() has proven to be a bottleneck in a number of
buffer acquisition heavy workloads. To some degree this has already
been alleviated by 5d7962c6, but it still can be quite a heavy
bottleneck. The problem is that in unfortunate usage patterns a
single StrategyGetBuffer() call will have to look at a large number of
buffers - in turn making it likely that the process will be put to
sleep while still holding the spinlock.
Replace most of the usage of the buffer_strategy_lock spinlock for the
clock sweep by a atomic nextVictimBuffer variable. That variable,
modulo NBuffers, is the current hand of the clock sweep. The buffer
clock-sweep then only needs to acquire the spinlock after a
wraparound. And even then only in the process that did the wrapping
around. That alleviates nearly all the contention on the relevant
spinlock, although significant contention on the cacheline can still
exist.
Reviewed-By: Robert Haas and Amit Kapila
Discussion: 20141010160020.GG6670@alap3.anarazel.de,
20141027133218.GA2639@awork2.anarazel.de
Previously, we used an lwlock that was held from the time we began
seeking a candidate buffer until the time when we found and pinned
one, which is disastrous for concurrency. Instead, use a spinlock
which is held just long enough to pop the freelist or advance the
clock sweep hand, and then released. If we need to advance the clock
sweep further, we reacquire the spinlock once per buffer.
This represents a significant increase in atomic operations around
buffer eviction, but it still wins on many workloads. On others, it
may result in no gain, or even cause a regression, unless the number
of buffer mapping locks is also increased. However, that seems like
material for a separate commit. We may also need to consider other
methods of mitigating contention on this spinlock, such as splitting
it into multiple locks or jumping the clock sweep hand more than one
buffer at a time, but those, too, seem like separate improvements.
Patch by me, inspired by a much larger patch from Amit Kapila.
Reviewed by Andres Freund.
This makes it possible to store lwlocks as part of some other data
structure in the main shared memory segment, or in a dynamic shared
memory segment. There is still a main LWLock array and this patch does
not move anything out of it, but it provides necessary infrastructure
for doing that in the future.
This change is likely to increase the size of LWLockPadded on some
platforms, especially 32-bit platforms where it was previously only
16 bytes.
Patch by me. Review by Andres Freund and KaiGai Kohei.
This provides a speedup of about 4X when NBuffers is large enough.
There is also a useful reduction in sinval traffic, since we
only do CacheInvalidateSmgr() once not once per fork.
Simon Riggs, reviewed and somewhat revised by Tom Lane
Commit 6d90eaaa89 added a hibernation mode
to the bgwriter to reduce the server's idle-power consumption. However,
its interaction with the detailed behavior of BgBufferSync's feedback
control loop wasn't very well thought out. That control loop depends
primarily on the rate of buffer allocation, not the rate of buffer
dirtying, so the hibernation mode has to be designed to operate only when
no new buffer allocations are happening. Also, the check for whether the
system is effectively idle was not quite right and would fail to detect
a constant low level of activity, thus allowing the bgwriter to go into
hibernation mode in a way that would let the cycle time vary quite a bit,
possibly further confusing the feedback loop. To fix, move the wakeup
support from MarkBufferDirty and SetBufferCommitInfoNeedsSave into
StrategyGetBuffer, and prevent the bgwriter from entering hibernation mode
unless no buffer allocations have happened recently.
In addition, fix the delaying logic to remove the problem of possibly not
responding to signals promptly, which was basically caused by trying to use
the process latch's is_set flag for multiple purposes. I can't prove it
but I'm suspicious that that hack was responsible for the intermittent
"postmaster does not shut down" failures we've been seeing in the buildfarm
lately. In any case it did nothing to improve the readability or
robustness of the code.
In passing, express the hibernation sleep time as a multiplier on
BgWriterDelay, not a constant. I'm not sure whether there's any value in
exposing the longer sleep time as an independently configurable setting,
but we can at least make it act like this for little extra code.
The contents of an unlogged table are WAL-logged; thus, they are not
available on standby servers and are truncated whenever the database
system enters recovery. Indexes on unlogged tables are also unlogged.
Unlogged GiST indexes are not currently supported.
This patch also removes buffer-usage statistics from the track_counts
output, since this (or the global server statistics) is deemed to be a better
interface to this information.
Itagaki Takahiro, reviewed by Euler Taveira de Oliveira.
Robert Haas