EXPLAIN (ANALYZE, SERIALIZE) allows collection of statistics about
the volume of data emitted by a query, as well as the time taken
to convert the data to the on-the-wire format. Previously there
was no way to investigate this without actually sending the data
to the client, in which case network transmission costs might
swamp what you wanted to see. In particular this feature allows
investigating the costs of de-TOASTing compressed or out-of-line
data during formatting.
Stepan Rutz and Matthias van de Meent,
reviewed by Tomas Vondra and myself
Discussion: https://postgr.es/m/ca0adb0e-fa4e-c37e-1cd7-91170b18cae1@gmx.de
After this change we have XLogCtl->logWriteResult and ->logFlushResult.
There's no functional change, other than the fact that the assignment
from shared memory to local is no longer done via struct assignment, but
instead using a macro that copies each member separately.
The current representation is inconvenient going forward; notably, we
would like to add a new member "Copy" (to keep track of the last
position copied into WAL buffers), so the symmetry between the values in
shared memory vs. those in local would be lost.
This also gives us freedom to later change the concurrency model for the
values in shared memory: we can make them use atomics instead of relying
on the info_lck spinlock.
Reviewed-by: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Discussion: https://postgr.es/m/202404031119.cd2kugjk2vho@alvherre.pgsql
If there aren't many bytes to process, the function call overhead
of the optimized implementation isn't worth taking, so instead we
inline a loop that consults pg_number_of_ones in that case. If
there are many bytes to process, we accept the function call
overhead because the optimized versions are likely to be faster.
The threshold at which we use the optimized implementation is set
to the smallest amount of data required to use special popcount
instructions.
Reviewed-by: Alvaro Herrera, Tom Lane
Discussion: https://postgr.es/m/20240402155301.GA2750455%40nathanxps13
Execute both freezing and pruning of tuples in the same
heap_page_prune() function, now called heap_page_prune_and_freeze(),
and emit a single WAL record containing all changes. That reduces the
overall amount of WAL generated.
This moves the freezing logic from vacuumlazy.c to the
heap_page_prune_and_freeze() function. The main difference in the
coding is that in vacuumlazy.c, we looked at the tuples after the
pruning had already happened, but in heap_page_prune_and_freeze() we
operate on the tuples before pruning. The heap_prepare_freeze_tuple()
function is now invoked after we have determined that a tuple is not
going to be pruned away.
VACUUM no longer needs to loop through the items on the page after
pruning. heap_page_prune_and_freeze() does all the work. It now
returns the list of dead offsets, including existing LP_DEAD items, to
the caller. Similarly it's now responsible for tracking 'all_visible',
'all_frozen', and 'hastup' on the caller's behalf.
Author: Melanie Plageman <melanieplageman@gmail.com>
Discussion: https://www.postgresql.org/message-id/20240330055710.kqg6ii2cdojsxgje@liskov
In preparation of freezing and counting tuples which are not
candidates for pruning, split heap_prune_record_unchanged() into
multiple functions, depending the kind of line pointer. That's not too
interesting right now, but makes the next commit smaller.
Recording the lowest soon-to-be prunable xid is one of the actions we
take for unchanged LP_NORMAL item pointers but not for others, so move
that to the new heap_prune_record_unchanged_lp_normal() function. The
next commit will add more actions to these functions.
Author: Melanie Plageman <melanieplageman@gmail.com>
Discussion: https://www.postgresql.org/message-id/20240330055710.kqg6ii2cdojsxgje@liskov
Bug in commit 53c2a97a9266: we failed to acquire the correct SLRU bank
lock when iterating to zero-out intermediate pages in predicate.c.
Rewrite the code block so that we follow the locking protocol correctly.
Also update an outdated comment in the same file -- SerialSLRULock
exists no more.
Reported-by: Alexander Lakhin <exclusion@gmail.com>
Reviewed-by: Dilip Kumar <dilipbalaut@gmail.com>
Discussion: https://postgr.es/m/2a25eaf4-a3a4-5fd1-6241-9d7c73142085@gmail.com
06c418e163 introduced pg_wal_replay_wait() procedure allowing to wait for
the particular LSN to be replayed on standby. The waiters were stored in
the flat array. Even though scanning small arrays is fast, that might be a
problem at scale (a lot of waiting processes).
This commit replaces the flat shared memory array with the pairing heap,
which holds the waiter with the least LSN at the top. This gives us O(log N)
complexity for both inserting and removing waiters.
Reported-by: Alvaro Herrera
Discussion: https://postgr.es/m/202404030658.hhj3vfxeyhft%40alvherre.pgsql
Project policy is to not leave global objects behind after a regress
test run. This was found as a result of the development of a patch
to make pg_regress detect such leftovers automatically, which in the
end was withdrawn due to issues with parallel runs.
Discussion: https://postgr.es/m/E1phvk7-000VAH-7k@gemulon.postgresql.org
We were directly copying the LSN locations while syncing the slots on the
standby. Now, it is possible that at some particular restart_lsn there are
some running xacts, which means if we start reading the WAL from that
location after promotion, we won't reach a consistent snapshot state at
that point. However, on the primary, we would have already been in a
consistent snapshot state at that restart_lsn so we would have just
serialized the existing snapshot.
To avoid this problem we will use the advance_slot functionality unless
the snapshot already exists at the synced restart_lsn location. This will
help us to ensure that snapbuilder/slot statuses are updated properly
without generating any changes. Note that the synced slot will remain as
RS_TEMPORARY till the decoding from corresponding restart_lsn can reach a
consistent snapshot state after which they will be marked as
RS_PERSISTENT.
Per buildfarm
Author: Hou Zhijie
Reviewed-by: Bertrand Drouvot, Shveta Malik, Bharath Rupireddy, Amit Kapila
Discussion: https://postgr.es/m/OS0PR01MB5716B3942AE49F3F725ACA92943B2@OS0PR01MB5716.jpnprd01.prod.outlook.com
This adds errcodes to a set of PANIC and FATAL errors in xlog.c
and relcache.c, which previously had no errcode at all set, in
order to make fleetwide analysis of errorlogs easier. There are
many more ereport/elogs left which could benefit from having an
errcode but this at least makes a dent in the issue.
Author: Nazir Bilal Yavuz <byavuz81@gmail.com>
Discussion: https://postgr.es/m/CAN55FZ1k8LgLEqncPGmz_fWnrobV6bjABOTH4tOWta6xNcPQig@mail.gmail.com
Presently, the archiver process restarts when an archive callback
ERRORs. To avoid this, archive module authors can use sigsetjmp(),
manage a memory context, etc., but that requires a lot of extra
code that will likely look roughly the same between modules. This
commit adds basic archive callback ERROR handling to pgarch.c so
that module authors won't ordinarily need to worry about this.
While this built-in handler attempts to clean up anything that an
archive module could conceivably have left behind, it is possible
that some modules are doing unexpected things that require
additional cleanup. Module authors should be sure to do any extra
required cleanup in a PG_CATCH block within the archiving callback.
The archiving callback is now called in a short-lived memory
context that the archiver process resets between invocations. If a
module requires longer-lived storage, it must maintain its own
memory context.
Thanks to these changes, the basic_archive module can be greatly
simplified.
Suggested-by: Andres Freund
Reviewed-by: Andres Freund, Yong Li
Discussion: https://postgr.es/m/20230217215624.GA3131134%40nathanxps13
Previously, when selecting the transaction to evict during logical
decoding, we check all transactions to find the largest
transaction. This could lead to a significant replication lag
especially in the case where there are many subtransactions.
This commit improves the eviction algorithm in ReorderBuffer using the
max-heap with transaction size as the key to efficiently find the
largest transaction.
The max-heap starts with empty. While the max-heap is empty, we don't
do anything for the max-heap when updating the memory
counter. Therefore, we get the largest transaction in O(N) time, where
N is the number of transactions including top-level transactions and
subtransactions.
We build the max-heap just before selecting the largest transactions
if the number of transactions being decoded is higher than the
threshold, MAX_HEAP_TXN_COUNT_THRESHOLD. After building the max-heap,
we also update the max-heap when updating the memory counter. The
intention is to efficiently find the largest transaction in O(1) time
instead of incurring the cost of memory counter updates (O(log
N)). Once the number of transactions got lower than the threshold, we
reset the max-heap.
The performance benchmark results showed significant speed up (more
than x30 speed up on my machine) in decoding a transaction with 100k
subtransactions, whereas there is no visible overhead in other cases.
Reviewed-by: Amit Kapila, Hayato Kuroda, Vignesh C, Ajin Cherian,
Tomas Vondra, Shubham Khanna, Peter Smith, Álvaro Herrera,
Euler Taveira
Discussion: https://postgr.es/m/CAD21AoAfKTgrBrLq96GcTv9d6k97zaQcDM-rxfKEt4GSe0qnaQ%40mail.gmail.com
Previously, binaryheap didn't support updating a key and removing a
node in an efficient way. For example, in order to remove a node from
the binaryheap, the caller had to pass the node's position within the
array that the binaryheap internally has. Removing a node from the
binaryheap is done in O(log n) but searching for the key's position is
done in O(n).
This commit adds a hash table to binaryheap in order to track the
position of each nodes in the binaryheap. That way, by using newly
added functions such as binaryheap_update_up() etc., both updating a
key and removing a node can be done in O(1) on an average and O(log n)
in worst case. This is known as the indexed binary heap. The caller
can specify to use the indexed binaryheap by passing indexed = true.
The current code does not use the new indexing logic, but it will be
used by an upcoming patch.
Reviewed-by: Vignesh C, Peter Smith, Hayato Kuroda, Ajin Cherian,
Tomas Vondra, Shubham Khanna
Discussion: https://postgr.es/m/CAD21AoDffo37RC-eUuyHJKVEr017V2YYDLyn1xF_00ofptWbkg%40mail.gmail.com
Since 66c0185a3, the planner is able to use Merge Append -> Unique to
implement UNION queries and each subquery is prompted to produce Paths
correctly sorted by the UNION's targetlist.
Here we remove some now redundant code which was zeroing the
tuple_fraction at the parent level. This will allow the planner to
consider cheap startup paths when planning the UNION's subqueries.
EXCEPT and INTERSECT set operations still have the tuple_fraction zeroed
in generate_nonunion_paths(). These operations currently always read
all of their subqueries' tuples.
Reported-by: Tom Lane
Discussion: https://postgr.es/m/3703023.1711654574@sss.pgh.pa.us
pg_wal_replay_wait() is to be used on standby and specifies waiting for
the specific WAL location to be replayed before starting the transaction.
This option is useful when the user makes some data changes on primary and
needs a guarantee to see these changes on standby.
The queue of waiters is stored in the shared memory array sorted by LSN.
During replay of WAL waiters whose LSNs are already replayed are deleted from
the shared memory array and woken up by setting of their latches.
pg_wal_replay_wait() needs to wait without any snapshot held. Otherwise,
the snapshot could prevent the replay of WAL records implying a kind of
self-deadlock. This is why it is only possible to implement
pg_wal_replay_wait() as a procedure working in a non-atomic context,
not a function.
Catversion is bumped.
Discussion: https://postgr.es/m/eb12f9b03851bb2583adab5df9579b4b%40postgrespro.ru
Author: Kartyshov Ivan, Alexander Korotkov
Reviewed-by: Michael Paquier, Peter Eisentraut, Dilip Kumar, Amit Kapila
Reviewed-by: Alexander Lakhin, Bharath Rupireddy, Euler Taveira
If temp tables have dependencies (such as sequences) then it's
possible for autovacuum's cleanup of orphan temp tables to deadlock
against an incoming backend that's trying to clean out the temp
namespace for its own use. That can happen because RemoveTempRelations'
performDeletion call can visit objects within the namespace in
an order different from the order in which a per-table deletion
will visit them.
To fix, observe that performDeletion will begin by taking an exclusive
lock on the temp namespace (even though it won't actually delete it).
So, if we can get a shared lock on the namespace, we can be sure we're
not running concurrently with RemoveTempRelations, while also not
conflicting with ordinary use of the namespace. This requires
introducing a conditional version of LockDatabaseObject, but that's no
big deal. (It's surprising we've got along without that this long.)
Report and patch by Mikhail Zhilin. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/c43ce028-2bc2-4865-9b89-3f706246eed5@postgrespro.ru
Make this code a little less confusing by defining a separate macro
that controls whether we'll use ResourceOwner facilities to track
the existence of a pg_hmac_ctx context.
The proximate reason to touch this is that since b8bff07da, we got
"unused function" warnings if building with older OpenSSL, because
the #if guards around the ResourceOwner wrapper function definitions
were different from those around the calls of those functions.
Pulling the ResourceOwner machinations outside of the #ifdef HAVE_xxx
guards fixes that and makes the code clearer too.
Discussion: https://postgr.es/m/1394271.1712016101@sss.pgh.pa.us
After installing the SIGINT handler in psql, SIGINT can no longer cancel
database reconnections. For instance, if the user starts a reconnection
and then needs to do some form of interaction (ie psql is polling),
there is no way to cancel the reconnection process currently.
Use PQconnectStartParams() in order to insert a cancel_pressed check
into the polling loop.
Tristan Partin, reviewed by Gurjeet Singh, Heikki Linnakangas, Jelte
Fennema-Nio, and me.
Discussion: http://postgr.es/m/D08WWCPVHKHN.3QELIKZJ2D9RZ@neon.tech
This is useful when connecting to a database asynchronously via
PQconnectStart(), since it handles deciding between poll() and
select(), and some of the required boilerplate.
Tristan Partin, reviewed by Gurjeet Singh, Heikki Linnakangas, Jelte
Fennema-Nio, and me.
Discussion: http://postgr.es/m/D08WWCPVHKHN.3QELIKZJ2D9RZ@neon.tech
Introduce an abstraction allowing relation data to be accessed as a
stream of buffers, with an implementation that is more efficient than
the equivalent sequence of ReadBuffer() calls.
Client code supplies a callback that can say which block number it wants
next, and then consumes individual buffers one at a time from the
stream. This division puts read_stream.c in control of how far ahead it
can see and allows it to read clusters of neighboring blocks with
StartReadBuffers(). It also issues POSIX_FADV_WILLNEED advice ahead of
time when random access is detected.
Other variants of I/O stream will be proposed in future work (for
example to support recovery, whose LsnReadQueue device in
xlogprefetcher.c is a distant cousin of this code and should eventually
be replaced by this), but this basic API is sufficient for many common
executor usage patterns involving predictable access to a single fork of
a single relation.
Several patches using this API are proposed separately.
This stream concept is loosely based on ideas from Andres Freund on how
we should pave the way for later work on asynchronous I/O.
Author: Thomas Munro <thomas.munro@gmail.com>
Author: Heikki Linnakangas <hlinnaka@iki.fi> (contributions)
Author: Melanie Plageman <melanieplageman@gmail.com> (contributions)
Suggested-by: Andres Freund <andres@anarazel.de>
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-by: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Nazir Bilal Yavuz <byavuz81@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Tested-by: Tomas Vondra <tomas.vondra@enterprisedb.com>
Discussion: https://postgr.es/m/CA+hUKGJkOiOCa+mag4BF+zHo7qo=o9CFheB8=g6uT5TUm2gkvA@mail.gmail.com
Break ReadBuffer() up into two steps. StartReadBuffers() and
WaitReadBuffers() give us two main advantages:
1. Multiple consecutive blocks can be read with one system call.
2. Advice (hints of future reads) can optionally be issued to the
kernel ahead of time.
The traditional ReadBuffer() function is now implemented in terms of
those functions, to avoid duplication.
A new GUC io_combine_limit is defined, and the functions for limiting
per-backend pin counts are made into public APIs. Those are provided
for use by callers of StartReadBuffers(), when deciding how many buffers
to read at once. The following commit will add a higher level mechanism
for doing that automatically with a practical interface.
With some more infrastructure in later work, StartReadBuffers() could
be extended to start real asynchronous I/O instead of just issuing
advice and leaving WaitReadBuffers() to do the work synchronously.
Author: Thomas Munro <thomas.munro@gmail.com>
Author: Andres Freund <andres@anarazel.de> (some optimization tweaks)
Reviewed-by: Melanie Plageman <melanieplageman@gmail.com>
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-by: Nazir Bilal Yavuz <byavuz81@gmail.com>
Reviewed-by: Dilip Kumar <dilipbalaut@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Tested-by: Tomas Vondra <tomas.vondra@enterprisedb.com>
Discussion: https://postgr.es/m/CA+hUKGJkOiOCa+mag4BF+zHo7qo=o9CFheB8=g6uT5TUm2gkvA@mail.gmail.com
Previously, we used a simple array for storing dead tuple IDs during
lazy vacuum, which had a number of problems:
* The array used a single allocation and so was limited to 1GB.
* The allocation was pessimistically sized according to table size.
* Lookup with binary search was slow because of poor CPU cache and
branch prediction behavior.
This commit replaces that array with the TID store from commit
30e144287a.
Since the backing radix tree makes small allocations as needed, the
1GB limit is now gone. Further, the total memory used is now often
smaller by an order of magnitude or more, depending on the
distribution of blocks and offsets. These two features should make
multiple rounds of heap scanning and index cleanup an extremely rare
event. TID lookup during index cleanup is also several times faster,
even more so when index order is correlated with heap tuple order.
Since there is no longer a predictable relationship between the number
of dead tuples vacuumed and the space taken up by their TIDs, the
number of tuples no longer provides any meaningful insights for users,
nor is the maximum number predictable. For that reason this commit
also changes to byte-based progress reporting, with the relevant
columns of pg_stat_progress_vacuum renamed accordingly to
max_dead_tuple_bytes and dead_tuple_bytes.
For parallel vacuum, both the TID store and supplemental information
specific to vacuum are shared among the parallel vacuum workers. As
with the previous array, we don't take any locks on TidStore during
parallel vacuum since writes are still only done by the leader
process.
Bump catalog version.
Reviewed-by: John Naylor, (in an earlier version) Dilip Kumar
Discussion: https://postgr.es/m/CAD21AoAfOZvmfR0j8VmZorZjL7RhTiQdVttNuC4W-Shdc2a-AA%40mail.gmail.com
66c0185a3 adjusted the UNION planner to request that union child queries
produce Paths correctly ordered to implement the UNION by way of
MergeAppend followed by Unique. The code there made a bad assumption
that if the root->parent_root->parse had setOperations set that the
query must be the child subquery of a set operation. That's not true
when it comes to planning a non-inlined CTE which is parented by a set
operation. This causes issues as the CTE's targetlist has no
requirement to match up to the SetOperationStmt's groupClauses
Fix this by adding a new parameter to both subquery_planner() and
grouping_planner() to explicitly pass the SetOperationStmt only when
planning set operation child subqueries.
Thank you to Tom Lane for helping to rationalize the decision on the
best function signature for subquery_planner().
Reported-by: Alexander Lakhin
Discussion: https://postgr.es/m/242fc7c6-a8aa-2daf-ac4c-0a231e2619c1@gmail.com
If we are building with openssl but USE_SSL_ENGINE didn't get set,
initialize_SSL's variable "pkey" is declared but used nowhere.
Apparently this combination hasn't been exercised in the buildfarm
before now, because I've not seen this warning before, even though
the code has been like this a long time. Move the declaration
to silence the warning (and remove its useless initialization).
Per buildfarm member sawshark. Back-patch to all supported branches.
Currently there is only one option, HEAP_PAGE_PRUNE_MARK_UNUSED_NOW
which replaces the old boolean argument, but upcoming patches will
introduce at least one more. Having a lot of boolean arguments makes
it hard to see at the call sites what the arguments mean, so prefer a
bitmask of options with human-readable names.
Author: Melanie Plageman <melanieplageman@gmail.com>
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Discussion: https://www.postgresql.org/message-id/20240401172219.fngjosaqdgqqvg4e@liskov
This patch allows pg_restore to wrap its commands into transaction
blocks, somewhat like --single-transaction, except that we commit
and start a new block after every N objects. Using this mode
with a size limit of 1000 or so objects greatly reduces the number
of transactions consumed by the restore, while preventing any
one transaction from taking enough locks to overrun the receiving
server's shared lock table.
(A value of 1000 works well with the default lock table size of
around 6400 locks. Higher --transaction-size values can be used
if one has increased the receiving server's lock table size.)
Excessive consumption of XIDs has been reported as a problem for
pg_upgrade in particular, but it could be bad for any restore; and the
change also reduces the number of fsyncs and amount of WAL generated,
so it should provide speed benefits too.
This patch does not try to make parallel workers batch the SQL
commands they issue. The trouble with doing that is that other
workers may need to see the objects a worker creates right away.
Possibly this can be improved later.
In this patch I have hard-wired pg_upgrade to use a transaction size
of 1000 divided by the number of parallel restore jobs allowed
(without that, we'd still be at risk of overrunning the shared lock
table). Perhaps there would be value in adding another pg_upgrade
option to allow user control of that, but I'm unsure that it's worth
the trouble; I think few users would use it, and any who did would see
not that much benefit compared to the default.
Patch by me, but the original idea to batch SQL commands during
restore is due to Robins Tharakan.
Discussion: https://postgr.es/m/a9f9376f1c3343a6bb319dce294e20ac@EX13D05UWC001.ant.amazon.com
Commit c0d5be5d6 caused pg_dump to create a separate BLOB metadata TOC
entry for each large object (blob), but it did not touch the ancient
decision to put all the blobs' data into a single "BLOBS" TOC entry.
This is bad for a few reasons: for databases with millions of blobs,
the TOC becomes unreasonably large, causing performance issues;
selective restore of just some blobs is quite impossible; and we
cannot parallelize either dump or restore of the blob data, since our
architecture for that relies on farming out whole TOC entries to
worker processes.
To improve matters, let's group multiple blobs into each blob metadata
TOC entry, and then make corresponding per-group blob data TOC entries.
Selective restore using pg_restore's -l/-L switches is then possible,
though only at the group level. (Perhaps we should provide a switch
to allow forcing one-blob-per-group for users who need precise
selective restore and don't have huge numbers of blobs. This patch
doesn't do that, instead just hard-wiring the maximum number of blobs
per entry at 1000.)
The blobs in a group must all have the same owner, since the TOC entry
format only allows one owner to be named. In this implementation
we also require them to all share the same ACL (grants); the archive
format wouldn't require that, but pg_dump's representation of
DumpableObjects does. It seems unlikely that either restriction
will be problematic for databases with huge numbers of blobs.
The metadata TOC entries now have a "desc" string of "BLOB METADATA",
and their "defn" string is just a newline-separated list of blob OIDs.
The restore code has to generate creation commands, ALTER OWNER
commands, and drop commands (for --clean mode) from that. We would
need special-case code for ALTER OWNER and drop in any case, so the
alternative of keeping the "defn" as directly executable SQL code
for creation wouldn't buy much, and it seems like it'd bloat the
archive to little purpose.
Since we require the blobs of a metadata group to share the same ACL,
we can furthermore store only one copy of that ACL, and then make
pg_restore regenerate the appropriate commands for each blob. This
saves space in the dump file not only by removing duplicative SQL
command strings, but by not needing a separate TOC entry for each
blob's ACL. In turn, that reduces client-side memory requirements for
handling many blobs.
ACL TOC entries that need this special processing are labeled as
"ACL"/"LARGE OBJECTS nnn..nnn". If we have a blob with a unique ACL,
continue to label it as "ACL"/"LARGE OBJECT nnn". We don't actually
have to make such a distinction, but it saves a few cycles during
restore for the easy case, and it seems like a good idea to not change
the TOC contents unnecessarily.
The data TOC entries ("BLOBS") are exactly the same as before,
except that now there can be more than one, so we'd better give them
identifying tag strings.
Also, commit c0d5be5d6 put the new BLOB metadata TOC entries into
SECTION_PRE_DATA, which perhaps is defensible in some ways, but
it's a rather odd choice considering that we go out of our way to
treat blobs as data. Moreover, because parallel restore handles
the PRE_DATA section serially, this means we'd only get part of the
parallelism speedup we could hope for. Move these entries into
SECTION_DATA, letting us parallelize the lo_create calls not just the
data loading when there are many blobs. Add dependencies to ensure
that we won't try to load data for a blob we've not yet created.
As this stands, we still generate a separate TOC entry for any comment
or security label attached to a blob. I feel comfortable in believing
that comments and security labels on blobs are rare, so this patch
should be enough to get most of the useful TOC compression for blobs.
We have to bump the archive file format version number, since existing
versions of pg_restore wouldn't know they need to do something special
for BLOB METADATA, plus they aren't going to work correctly with
multiple BLOBS entries or multiple-large-object ACL entries.
The directory and tar-file format handlers need some work
for multiple BLOBS entries: they used to hard-wire the file name
as "blobs.toc", which is replaced here with "blobs_<dumpid>.toc".
The 002_pg_dump.pl test script also knows about that and requires
minor updates. (I had to drop the test for manually-compressed
blobs.toc files with LZ4, because lz4's obtuse command line
design requires explicit specification of the output file name
which seems impractical here. I don't think we're losing any
useful test coverage thereby; that test stanza seems completely
duplicative with the gzip and zstd cases anyway.)
In passing, centralize management of the lo_buf used to hold data
while restoring blobs. The code previously had each format handler
create lo_buf, which seems rather pointless given that the format
handlers all make it the same way. Moreover, the format handlers
never use lo_buf directly, making this setup a failure from a
separation-of-concerns standpoint. Let's move the responsibility into
pg_backup_archiver.c, which is the only module concerned with lo_buf.
The reason to do this in this patch is that it allows a centralized
fix for the now-false assumption that we never restore blobs in
parallel. Also, get rid of dead code in DropLOIfExists: it's been a
long time since we had any need to be able to restore to a pre-9.0
server.
Discussion: https://postgr.es/m/a9f9376f1c3343a6bb319dce294e20ac@EX13D05UWC001.ant.amazon.com
The "pltargs" variable wasn't marked volatile, which makes it unsafe
to change its value within the PG_TRY block. It looks like the worst
outcome would be to fail to release a refcount on Py_None during an
(improbable) error exit, which would likely go unnoticed in the field.
Still, it's a bug. A one-liner fix could be to mark pltargs volatile,
but on the whole it seems cleaner to arrange things so that we don't
change its value within PG_TRY.
Per report from Xing Guo. This has been there for quite awhile,
so back-patch to all supported branches.
Discussion: https://postgr.es/m/CACpMh+DLrk=fDv07MNpBT4J413fDAm+gmMXgi8cjPONE+jvzuw@mail.gmail.com
Various error paths did not release resources before returning.
While it's likely that the program would just exit shortly later,
none of the functions in question have summary exit(1) calls,
so they should not be assuming that.
Ranier Vilela and Tom Lane, per reports from Coverity
Discussion: https://postgr.es/m/CAEudQAr2_SZFxB4kXJiL4+2UaNZxUk5UBJtj0oXyJYMGZu-03g@mail.gmail.com
Handle dead branches of aborted HOT chains outside heap_prune_chain()
as a separate phase. This simplifies the logic in heap_prune_chain(),
as well as allowing us to clean up more RECENTLY_DEAD -> DEAD chains.
To accomplish this efficiently, partition tuples into HOT and non-HOT
while first collecting visibility information for each tuple in
heap_page_prune(). Then call heap_prune_chain() only on potential
chain members. Then mop up the leftover HOT tuples afterwards.
As part of this, keep track of which items on page have already been
processed, in 'processed' array. This replaces the 'marked' array
which was only set for tuples marked for removal or redirection. The
'processed' array is updated also for items that are left unchanged,
when we conclude that an item can be left unchanged. At the end of
pruning, every item on the page should be marked as processed in the
array; an assertion is added for that.
Author: Melanie Plageman <melanieplageman@gmail.com>
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Discussion: https://www.postgresql.org/message-id/20240330055710.kqg6ii2cdojsxgje@liskov
Keep track of the number of deleted tuples in PruneState and record this
information when recording a tuple dead, unused or redirected. This
removes a special case from the traversal and chain processing logic as
well as setting a precedent of recording the impact of prune actions in
the record functions themselves. This paradigm will be used in future
commits which move tracking of additional statistics on pruning actions
from lazy_scan_prune() to heap_prune_chain().
Simplify heap_prune_chain()'s chain traversal logic by handling each
case explicitly. That is, do not attempt to share code when processing
different types of chains. For each category of chain, process it
specifically and procedurally: first handling the root, then any
intervening tuples, and, finally, the end of the chain.
While we are at it, add a few new comments to heap_prune_chain()
clarifying some special cases involving RECENTLY_DEAD tuples.
Author: Melanie Plageman <melanieplageman@gmail.com>
Discussion: https://www.postgresql.org/message-id/20240330055710.kqg6ii2cdojsxgje@liskov
Pass 'page', 'blockno' and 'maxoff' to heap_prune_chain() as
arguments, so that it doesn't need to fetch them from the buffer. This
saves a few cycles per chain.
Remove the "if (off_loc != NULL)" checks, and require the caller to
pass a non-NULL 'off_loc'. Pass a pointer to a dummy local variable
when it's not needed. Those checks are cheap, but it's still better to
avoid them in the per-chain loops when we can do so easily.
The CPU time saving from these changes are hardly measurable, but
fewer instructions is good anyway, so why not. I spotted the potential
for these while reviewing Melanie Plageman's patch set to combine
prune and freeze records.
Discussion: https://www.postgresql.org/message-id/CAAKRu_abm2tHhrc0QSQa%3D%3DsHe%3DVA1%3Doz1dJMQYUOKuHmu%2B9Xrg%40mail.gmail.com
This commit adds a new COPY option LOG_VERBOSITY, which controls the
amount of messages emitted during processing. Valid values are
'default' and 'verbose'.
This is currently used in COPY FROM when ON_ERROR option is set to
ignore. If 'verbose' is specified, a NOTICE message is emitted for
each discarded row, providing additional information such as line
number, column name, and the malformed value. This helps users to
identify problematic rows that failed to load.
Author: Bharath Rupireddy
Reviewed-by: Michael Paquier, Atsushi Torikoshi, Masahiko Sawada
Discussion: https://www.postgresql.org/message-id/CALj2ACUk700cYhx1ATRQyRw-fBM%2BaRo6auRAitKGff7XNmYfqQ%40mail.gmail.com
After encountering the NUL terminator, the word-at-a-time loop exits
and we must hash the remaining bytes. Previously we calculated the
terminator's position and re-loaded the remaining bytes from the input
string. We already have all the data we need in a register, so let's
just mask off the bytes we need and hash them immediately. The mask can
be cheaply computed without knowing the terminator's position. We still
need that position for the length calculation, but the CPU can now
do that in parallel with other work, shortening the dependency chain.
Ants Aasma and John Naylor
Discussion: https://postgr.es/m/CANwKhkP7pCiW_5fAswLhs71-JKGEz1c1%2BPC0a_w1fwY4iGMqUA%40mail.gmail.com
Previously, the executor did index insert unconditionally after calling
table AM interface methods tuple_insert() and multi_insert(). This commit
introduces the new parameter insert_indexes for these two methods. Setting
'*insert_indexes' to true saves the current logic. Setting it to false
indicates that table AM cares about index inserts itself and doesn't want the
caller to do that.
Discussion: https://postgr.es/m/CAPpHfdurb9ycV8udYqM%3Do0sPS66PJ4RCBM1g-bBpvzUfogY0EA%40mail.gmail.com
Reviewed-by: Pavel Borisov, Matthias van de Meent, Mark Dilger
Jeff Davis