after actually removing one, so that if we can't remove segments because
WAL archiving is lagging behind, we don't unnecessarily forbid streaming
the old not-yet-archived segments that are still perfectly valid. Per
suggestion from Fujii Masao.
doesn't take into account how far the WAL senders are. This way a hung
WAL sender doesn't prevent old WAL segments from being recycled/removed
in the primary, ultimately causing the disk to fill up. Instead add
standby_keep_segments setting to control how many old WAL segments are
kept in the primary. This also makes it more reliable to use streaming
replication without WAL archiving, assuming that you set
standby_keep_segments high enough.
recovery. We might want to relax this in the future, but ThisTimeLineID
isn't currently correct in backends during recovery, so the filename
returned was wrong.
This allows us to see what mode the server is in before it starts to
perform actions that can block or hang. Otherwise server messages
may not appear until after messages that say FATAL the database
server is starting up.
The error message now makes explicit reference to the GUC that must be changed
to fix the problem, using wording suggested by Tom Lane. Along the way,
rename the GUC from MaxWalSenders to max_wal_senders for consistency and
grep-ability.
pg_xlog directory. This is essential for replaying WAL records that
were streamed from the master, after a standby server restart.
If a corrupt record is seen in a file restored from the archive or
streamed from the master, log it as a WARNING and keep retrying. If the
corruption is permanent, and not just a glitch in the whatever copies the
files to the archive or a network error not caught by CRC checks in TCP
for example, we will keep retrying and logging the WARNING indefinitely.
But that's better than shutting down completely, the standby is still
useful for running read-only queries. In PITR the recovery ends at such a
corrupt record, which is a bit questionable, but that's the behavior we
had in previous releases and we don't feel like chaning it now. It does
make sense for tools like pg_standby.
example to 'on or 'off' rather than 'true' or 'false', as shown
in docs. Add restartpoint_command. Add section header for recovery
target parameters, matching docs.
WAL record for btree delete contains a list of tids, even when backup
blocks are present. We follow the tids to their heap tuples, taking
care to follow LP_REDIRECT tuples. We ignore LP_DEAD tuples on the
understanding that they will always have xmin/xmax earlier than any
LP_NORMAL tuples referred to by killed index tuples. Iff all tuples
are LP_DEAD we return InvalidTransactionId. The heap relfilenode is
added to the WAL record, requiring API changes to pass down the heap
Relation. XLOG_PAGE_MAGIC updated.
present since 8.0 was never fully meaningful, since two recovery targets
cannot be specified. Refactor recovery target type to make this change
and associated code easier to understand. No change in function.
Bug report arising from internal support question.
field into WAL record and reset it from there, rather than using
FrozenTransactionId which can lead to some corner case bugs.
Problem report and suggested route to a fix from Heikki, details by me.
in recovery_end_command, it always came out as 0 because InRedo was
cleared before recovery_end_command was executed. Also, always take
ControlFileLock when reading checkpoint location for %r.
The recovery_end_command bug and the missing locking was present in 8.4
as well, that part of this patch will be backported separately.
instead of an exclusive lock.
The change is almost for code cleanup. Since there seems to be no
performance benefits from it, backports should not be needed.
Fujii Masao
Add some checks that seem logically necessary, in particular let's make
real sure that HS slave sessions cannot create temp tables. (If they did
they would think that temp tables belonging to the master's session with
the same BackendId were theirs. We *must* not allow myTempNamespace to
become set in a slave session.)
Change setval() and nextval() so that they are only allowed on temp sequences
in a read-only transaction. This seems consistent with what we allow for
table modifications in read-only transactions. Since an HS slave can't have a
temp sequence, this also provides a nicer cure for the setval PANIC reported
by Erik Rijkers.
Make the error messages more uniform, and have them mention the specific
command being complained of. This seems worth the trifling amount of extra
code, since people are likely to see such messages a lot more than before.
enabled. Bypassing the kernel cache is counter-productive in that case,
because the archiver/walsender process will read from the WAL file
soon after it's written, and if it's not cached the read will cause
a physical read, eating I/O bandwidth available on the WAL drive.
Also, walreceiver process does unaligned writes, so disable O_DIRECT
in walreceiver process for that reason too.
segment of XLOG_BACKUP_END record even if the the record is placed
at a segment boundary. Furthermore the previous implementation could
return nonexistent segment file name when the boundary is in segments
that has "FE" suffix; We never use segments with "FF" suffix.
Backpatch to 8.0, where hot backup was introduced.
Reported by Fujii Masao.
all the values derived from oldestXid, not just that field. Brain fade in
one of my patches associated with flat file removal, exposed by a report
from Fujii Masao.
With this change, xidVacLimit should always be valid, so remove a couple of
bits of complexity associated with the previous assumption that sometimes
it wouldn't get set right away.
In addition, add support for a "payload" string to be passed along with
each notify event.
This implementation should be significantly more efficient than the old one,
and is also more compatible with Hot Standby usage. There is not yet any
facility for HS slaves to receive notifications generated on the master,
although such a thing is possible in future.
Joachim Wieland, reviewed by Jeff Davis; also hacked on by me.
The purpose of this change is to eliminate the need for every caller
of SearchSysCache, SearchSysCacheCopy, SearchSysCacheExists,
GetSysCacheOid, and SearchSysCacheList to know the maximum number
of allowable keys for a syscache entry (currently 4). This will
make it far easier to increase the maximum number of keys in a
future release should we choose to do so, and it makes the code
shorter, too.
Design and review by Tom Lane.
where a database has a non-default tablespaceid. Pass thru MyDatabaseId
and MyDatabaseTableSpace to allow file path to be re-created in
standby and correct invalidation to take place in all cases.
Update and rework xact_commit_desc() debug messages.
Bug report from Tom by code inspection. Fix by me.
echo all the recovery.conf options. Don't emit the "initializing
recovery connections" message, which doesn't mean anything to a user.
Remove the "starting archive recovery" message and replace the
"automatic recovery in progress" message with a more informative message
saying whether the server is doing PITR, normal archive recovery, or
standby mode.
a partial WAL file, assume it's because the file is just being copied to
the archive and treat it the same as "file not found" in standby mode.
pg_standby has a similar check, so it seems reasonable to have the same
level of protection in the built-in standby mode.
several places, but for now only GIN uses it during index creation.
Using self-balanced tree greatly speeds up index creation in corner cases
with preordered data.
restoring from archive, the last WAL segment is not necessarily open at
the end of recovery. Fix assertion that assumed that.
Fujii Masao, fixing the assertion failure reported by Martin Pihlak.
Move rd_targblock, rd_fsm_nblocks, and rd_vm_nblocks from relcache to the smgr
relation entries, so that they will get reset to InvalidBlockNumber whenever
an smgr-level flush happens. Because we now send smgr invalidation messages
immediately (not at end of transaction) when a relation truncation occurs,
this ensures that other backends will reset their values before they next
access the relation. We no longer need the unreliable assumption that a
VACUUM that's doing a truncation will hold its AccessExclusive lock until
commit --- in fact, we can intentionally release that lock as soon as we've
completed the truncation. This patch therefore reverts (most of) Alvaro's
patch of 2009-11-10, as well as my marginal hacking on it yesterday. We can
also get rid of assorted no-longer-needed relcache flushes, which are far more
expensive than an smgr flush because they kill a lot more state.
In passing this patch fixes smgr_redo's failure to perform visibility-map
truncation, and cleans up some rather dubious assumptions in freespace.c and
visibilitymap.c about when rd_fsm_nblocks and rd_vm_nblocks can be out of
date.
recovery. It's zeroed out whenever a checkpoint is written, so the only
scenario where the removed code did anything is when you kill archive
recovery, remove recovery.conf, and start up the server, so that it goes
into crash recovery instead. That's a "don't do that" scenario, but it
seems better to not clear minRecoveryPoint but instead update it like we
do in archive recovery, which is what will now happen.
VACUUM FULL INPLACE), along with a boatload of subsidiary code and complexity.
Per discussion, the use case for this method of vacuuming is no longer large
enough to justify maintaining it; not to mention that we don't wish to invest
the work that would be needed to make it play nicely with Hot Standby.
Aside from the code directly related to old-style VACUUM FULL, this commit
removes support for certain WAL record types that could only be generated
within VACUUM FULL, redirect-pointer removal in heap_page_prune, and
nontransactional generation of cache invalidation sinval messages (the last
being the sticking point for Hot Standby).
We still have to retain all code that copes with finding HEAP_MOVED_OFF and
HEAP_MOVED_IN flag bits on existing tuples. This can't be removed as long
as we want to support in-place update from pre-9.0 databases.
of shared or nailed system catalogs. This has two key benefits:
* The new CLUSTER-based VACUUM FULL can be applied safely to all catalogs.
* We no longer have to use an unsafe reindex-in-place approach for reindexing
shared catalogs.
CLUSTER on nailed catalogs now works too, although I left it disabled on
shared catalogs because the resulting pg_index.indisclustered update would
only be visible in one database.
Since reindexing shared system catalogs is now fully transactional and
crash-safe, the former special cases in REINDEX behavior have been removed;
shared catalogs are treated the same as non-shared.
This commit does not do anything about the recently-discussed problem of
deadlocks between VACUUM FULL/CLUSTER on a system catalog and other
concurrent queries; will address that in a separate patch. As a stopgap,
parallel_schedule has been tweaked to run vacuum.sql by itself, to avoid
such failures during the regression tests.
of old and new toast tables can be done either at the logical level (by
swapping the heaps' reltoastrelid links) or at the physical level (by swapping
the relfilenodes of the toast tables and their indexes). This is necessary
infrastructure for upcoming changes to support CLUSTER/VAC FULL on shared
system catalogs, where we cannot change reltoastrelid. The physical swap
saves a few catalog updates too.
We unfortunately have to keep the logical-level swap logic because in some
cases we will be adding or deleting a toast table, so there's no possibility
of a physical swap. However, that only happens as a consequence of schema
changes in the table, which we do not need to support for system catalogs,
so such cases aren't an obstacle for that.
In passing, refactor the cluster support functions a little bit to eliminate
unnecessarily-duplicated code; and fix the problem that while CLUSTER had
been taught to rename the final toast table at need, ALTER TABLE had not.
heap_sync() to the callers, because heap_sync() is sometimes called even
if the operation itself is WAL-logged. This eliminates the bogus unlogged
records from CLUSTER that Simon Riggs reported, patch by Fujii Masao.
false positives during Hot Standby conflict processing. Simple
patch to enhance conflict processing, following previous discussions.
Controlled by parameter minimize_standby_conflicts = on | off, with
default off allows measurement of performance impact to see whether
it should be set on all the time.
records for heap and btree. Minor change, mostly API changes to
pass through the required values. This is a simple change though
also provides the refactoring required for further enhancements
to conflict processing using the relOid. Changes only have effect
during Hot Standby.
LogwrtRqst.Write can be set to non-existent FF log segment, we mustn't
try to send that in XLogSend().
Also fix similar bug in ReadRecord(), which I just introduced in the
ReadRecord() refactoring patch.
restore_command, if the connection to the primary server is lost. This
ensures that the standby can recover automatically, if the connection is
lost for a long time and standby falls behind so much that the required
WAL segments have been archived and deleted in the master.
This also makes standby_mode useful without streaming replication; the
server will keep retrying restore_command every few seconds until the
trigger file is found. That's the same basic functionality pg_standby
offers, but without the bells and whistles.
To implement that, refactor the ReadRecord/FetchRecord functions. The
FetchRecord() function introduced in the original streaming replication
patch is removed, and all the retry logic is now in a new function called
XLogReadPage(). XLogReadPage() is now responsible for executing
restore_command, launching walreceiver, and waiting for new WAL to arrive
from primary, as required.
This also changes the life cycle of walreceiver. When launched, it now only
tries to connect to the master once, and exits if the connection fails, or
is lost during streaming for any reason. The startup process detects the
death, and re-launches walreceiver if necessary.
AbortTransaction or AbortSubTransaction, when trying to clean up after an
error that prevented (sub)transaction start from completing:
* access to TopTransactionResourceOwner that might not exist
* assert failure in AtEOXact_GUC, if AtStart_GUC not called yet
* assert failure or core dump in AfterTriggerEndSubXact, if
AfterTriggerBeginSubXact not called yet
Per testing by injecting elog(ERROR) at successive steps in StartTransaction
and StartSubTransaction. It's not clear whether all of these cases could
really occur in the field, but at least one of them is easily exposed by
simple stress testing, as per my accidental discovery yesterday.
woken by alarm we send SIGUSR1 to all backends requesting that they
check to see if they are blocking Startup process. If so, they throw
ERROR/FATAL as for other conflict resolutions. Deadlock stop gap
removed. max_standby_delay = -1 option removed to prevent deadlock.
Attributes can now have options, just as relations and tablespaces do, and
the reloptions code is used to parse, validate, and store them. For
simplicity and because these options are not performance critical, we store
them in a separate cache rather than the main relcache.
Thanks to Alex Hunsaker for the review.
that would've been WAL-logged if archiving was enabled. If we encounter
such records in archive recovery anyway, we know that some data is
missing from the log. A WARNING is emitted in that case.
Original patch by Fujii Masao, with changes by me.
Conflict reason is passed through directly to the backend, so we can
take decisions about the effect of the conflict based upon the local
state. No specific changes, as yet, though this prepares for later work.
CancelVirtualTransaction() sends signals while holding ProcArrayLock.
Introduce errdetail_abort() to give message detail explaining that the
abort was caused by conflict processing. Remove CONFLICT_MODE states
in favour of using PROCSIG_RECOVERY_CONFLICT states directly, for clarity.
This includes two new kinds of postmaster processes, walsenders and
walreceiver. Walreceiver is responsible for connecting to the primary server
and streaming WAL to disk, while walsender runs in the primary server and
streams WAL from disk to the client.
Documentation still needs work, but the basics are there. We will probably
pull the replication section to a new chapter later on, as well as the
sections describing file-based replication. But let's do that as a separate
patch, so that it's easier to see what has been added/changed. This patch
also adds a new section to the chapter about FE/BE protocol, documenting the
protocol used by walsender/walreceivxer.
Bump catalog version because of two new functions,
pg_last_xlog_receive_location() and pg_last_xlog_replay_location(), for
monitoring the progress of replication.
Fujii Masao, with additional hacking by me
of this are to centralise the conflict code to allow further change,
as well as to allow passing through the full reason for the conflict
through to the conflicting backends. Backend state alters how we
can handle different types of conflict so this is now required.
As originally suggested by Heikki, no longer optional.
Previously, fastgetattr() and heap_getattr() tested their fourth argument
against a null pointer, but any attempt to use them with a literal-NULL
fourth argument evaluated to *(void *)0, resulting in a compiler error.
Remove these NULL tests to avoid leading future readers of this code to
believe that this has a chance of working. Also clean up related legacy
code in nocachegetattr(), heap_getsysattr(), and nocache_index_getattr().
The new coding standard is that any code which calls a getattr-type
function or macro which takes an isnull argument MUST pass a valid
boolean pointer. Per discussion with Bruce Momjian, Tom Lane, Alvaro
Herrera.
deletion, so that we attempt to unlink the correct filepath. unlink()
errors are ignorable there, so lack of a DatabasePath initialization step
did not cause visible problems until a related bug showed up on Solaris.
Code refactored from xact_redo_commit() to
ProcessCommittedInvalidationMessages() in inval.c. Recovery may replay
shared invalidation messages for many databases, so we cannot
SetDatabasePath() once as we do in normal backends. Read the databaseid
from the shared invalidation messages, then set DatabasePath
temporarily before calling RelationCacheInitFileInvalidate().
Problem report by Robert Treat, analysis and fix by me.
we're not going to support that anymore.
I did keep the 64-bit-CRC-with-32-bit-arithmetic code, since it has a
performance excuse to live. It's a bit moot since that's all ifdef'd
out, of course.
This patch only supports seq_page_cost and random_page_cost as parameters,
but it provides the infrastructure to scalably support many more.
In particular, we may want to add support for effective_io_concurrency,
but I'm leaving that as future work for now.
Thanks to Tom Lane for design help and Alvaro Herrera for the review.
recovery instead of reading the backup history file. This is more robust,
as it stops you from prematurely starting up an inconsisten cluster if the
backup history file is lost for some reason, or if the base backup was
never finished with pg_stop_backup().
This also paves the way for a simpler streaming replication patch, which
doesn't need to care about backup history files anymore.
The backup history file is still created and archived as before, but it's
not used by the system anymore. It's just for informational purposes now.
Bump PG_CONTROL_VERSION as the location of the backup startpoint is now
written to a new field in pg_control, and catversion because initdb is
required
Original patch by Fujii Masao per Simon's idea, with further fixes by me.
indexscans would do the wrong thing if index_rescan() was called with a
NULL instead of a new set of scankeys and the index was DESC order,
because sk_strategy would not get flipped a second time. I think
that those provisions for a NULL argument are dead code now as far as the
core backend goes, but possibly somebody somewhere is still using it.
In any case, this refactoring seems clearer, and it's definitely shorter.
to be just a minor extension of the previous patch that made "x IS NULL"
indexable, because we can treat the IS NOT NULL condition as if it were
"x < NULL" or "x > NULL" (depending on the index's NULLS FIRST/LAST option),
just like IS NULL is treated like "x = NULL". Aside from any possible
usefulness in its own right, this is an important improvement for
index-optimized MAX/MIN aggregates: it is now reliably possible to get
a column's min or max value cheaply, even when there are a lot of nulls
cluttering the interesting end of the index.
This is more in keeping with modern practice, and is a first step towards
porting to Win64 (which has sizeof(pointer) > sizeof(long)).
Tsutomu Yamada, Magnus Hagander, Tom Lane
index page split. This would result in index corruption, or even more likely
an error during WAL replay, if we were unlucky enough to crash during
end-of-recovery cleanup after having completed an incomplete GIST insertion.
Yoichi Hirai
Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record.
New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far.
This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required.
Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit.
Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
an allegedly immutable index function. It was previously recognized that
we had to prevent such a function from executing SET/RESET ROLE/SESSION
AUTHORIZATION, or it could trivially obtain the privileges of the session
user. However, since there is in general no privilege checking for changes
of session-local state, it is also possible for such a function to change
settings in a way that might subvert later operations in the same session.
Examples include changing search_path to cause an unexpected function to
be called, or replacing an existing prepared statement with another one
that will execute a function of the attacker's choosing.
The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against
these threats, which are the same places previously deemed to need protection
against the SET ROLE issue. GUC changes are still allowed, since there are
many useful cases for that, but we prevent security problems by forcing a
rollback of any GUC change after completing the operation. Other cases are
handled by throwing an error if any change is attempted; these include temp
table creation, closing a cursor, and creating or deleting a prepared
statement. (In 7.4, the infrastructure to roll back GUC changes doesn't
exist, so we settle for rejecting changes of "search_path" in these contexts.)
Original report and patch by Gurjeet Singh, additional analysis by
Tom Lane.
Security: CVE-2009-4136
support any indexable commutative operator, not just equality. Two rows
violate the exclusion constraint if "row1.col OP row2.col" is TRUE for
each of the columns in the constraint.
Jeff Davis, reviewed by Robert Haas
be part of multixacts, so allocate a slot for each prepared transaction in
the "oldest member" array in multixact.c. On PREPARE TRANSACTION, transfer
the oldest member value from the current backends slot to the prepared xact
slot. Also save and recover the value from the 2pc state file.
The symptom of the bug was that after a transaction prepared, a shared lock
still held by the prepared transaction was sometimes ignored by other
transactions.
Fix back to 8.1, where both 2PC and multixact were introduced.
User-defined consistent functions believes the check array
contains at least one true element which was not a true for
scanning pending list.
Per report from Yury Don <yura@vpcit.ru>
that it can scribble on scan->xs_ctup.t_self while following HOT chains,
so we can't rely on that to stay valid between hashgettuple() calls.
Introduce a private variable in HashScanOpaque, instead.
hash indexes keep entries sorted by hash value. First, the original plans for
concurrency assumed that insertions would happen only at the end of a page,
which is no longer true; this could cause scans to transiently fail to find
index entries in the presence of concurrent insertions. We can compensate
by teaching scans to re-find their position after re-acquiring read locks.
Second, neither the bucket split nor the bucket compaction logic had been
fixed to preserve hashvalue ordering, so application of either of those
processes could lead to permanent corruption of an index, in the sense
that searches might fail to find entries that are present.
This patch fixes the split and compaction logic to preserve hashvalue
ordering, but it cannot do anything about pre-existing corruption. We will
need to recommend reindexing all hash indexes in the 8.4.2 release notes.
To buy back the performance loss hereby induced in split and compaction,
fix them to use PageIndexMultiDelete instead of retail PageIndexDelete
operations. We might later want to do something with qsort'ing the
page contents rather than doing a binary search for each insertion,
but that seemed more invasive than I cared to risk in a back-patch.
Per bug #5157 from Jeff Janes and subsequent investigation.
friends). This code has all been ifdef'd out for many years, and doesn't
seem to have any prospect of becoming any more useful in the future.
EXPLAIN ANALYZE is what people use in practice, and I think if we did want
process-wide counters we'd be more likely to put in dtrace events for that
than try to resurrect this code. Get rid of it so as to have one less detail
to worry about while refactoring execMain.c.
tuple size limit. Improve the error message for index-tuple-too-large
so that it includes the actual size, the limit, and the index name.
Sync with the btree occurrences of the same error.
Back-patch to 8.4 because it appears that the out-of-sync problem
is occurring in the field.
Teodor and Tom
In practice these mistakes were always masked when full_page_writes was on,
because XLogInsert would always choose to log the full page, and then
ginRedoInsertListPage wouldn't try to do anything. But with full_page_writes
off a WAL replay failure was certain.
The GIN_INSERT_LISTPAGE record type could probably be eliminated entirely
in favor of using XLOG_HEAP_NEWPAGE, but I refrained from doing that now
since it would have required a significantly more invasive patch.
In passing do a little bit of code cleanup, including making the accounting
for free space on GIN list pages more precise. (This wasn't a bug as the
errors were always in the conservative direction.)
Per report from Simon. Back-patch to 8.4 which contains the identical code.
of checkpoint. Although the checkpoint has been written to WAL at that point
already, so that all data is safe, and we'll retry removing the WAL segment at
the next checkpoint, if such a failure persists we won't be able to remove any
other old WAL segments either and will eventually run out of disk space. It's
better to treat the failure as non-fatal, and move on to clean any other WAL
segment and continue with any other end-of-checkpoint cleanup.
We don't normally expect any such failures, but on Windows it can happen with
some anti-virus or backup software that lock files without FILE_SHARE_DELETE
flag.
Also, the loop in pgrename() to retry when the file is locked was broken. If a
file is locked on Windows, you get ERROR_SHARE_VIOLATION, not
ERROR_ACCESS_DENIED, at least on modern versions. Fix that, although I left
the check for ERROR_ACCESS_DENIED in there as well (presumably it was correct
in some environment), and added ERROR_LOCK_VIOLATION to be consistent with
similar checks in pgwin32_open(). Reduce the timeout on the loop from 30s to
10s, on the grounds that since it's been broken, we've effectively had a
timeout of 0s and no-one has complained, so a smaller timeout is actually
closer to the old behavior. A longer timeout would mean that if recycling a
WAL file fails because it's locked for some reason, InstallXLogFileSegment()
will hold ControlFileLock for longer, potentially blocking other backends, so
a long timeout isn't totally harmless.
While we're at it, set errno correctly in pgrename().
Backpatch to 8.2, which is the oldest version supported on Windows. The xlog.c
changes would make sense on other platforms and thus on older versions as
well, but since there's no such locking issues on other platforms, it's not
worth it.
file handle on it, the file goes into "pending deletion" state where it
still shows up in directory listing, but isn't accessible otherwise. That
confuses RemoveOldXLogFiles(), making it think that the file hasn't been
archived yet, while it actually was, and it was deleted along with the .done
file.
Fix that by renaming the file with ".deleted" extension before deleting it.
Also check the return value of rename() and unlink(), so that if the removal
fails for any reason (e.g another process is holding the file locked), we
don't delete the .done file until the WAL file is really gone.
Backpatch to 8.2, which is the oldest version supported on Windows.
own database's datfrozenxid, if the current value is old enough to be
forcing autovacuums or warning messages. This ensures that a bogus
value is replaced as soon as possible. Per a comment from Heikki.
Recent commits have removed the various uses it was supporting. It was a
performance bottleneck, according to bug report #4919 by Lauris Ulmanis; seems
it slowed down user creation after a billion users.
XID) in checkpoint records. This eliminates the need to recompute the value
from scratch during database startup, which is one of the two remaining
reasons for the flatfile code to exist. It should also simplify life for
hot-standby operation.
To avoid bloating the checkpoint records unreasonably, I switched from
tracking the oldest database by name to tracking it by OID. This turns
out to save cycles in general (everywhere but the warning-generating
paths, which we hardly care about) and also helps us deal with the case
that the oldest database got dropped instead of being vacuumed. The prior
coding might go for a long time without updating the wrap limit in that case,
which is bad because it might result in a lot of useless autovacuum activity.
In the original coding, setting a single reloption would cause default
values to be used for all the other reloptions. This is a problem
particularly for autovacuum reloptions.
Itagaki Takahiro
was incorrectly initialized with timeline ID 0. That rendered the WAL page
unrecoverable, making a subsequent archive recovery stop at that point.
ThisTimeLineID needs to be initialized before calling AdvanceXLInsertBuffer().
This fixes bug #5011 reported by James Bardin. Backpatch to 8.4, as the bug
was introduced by the changes to use of bgwriter for writing the
end-of-archive-recovery checkpoint. Patch by Tom Lane.
"all tuples visible" flag in heap page headers. The flag update *must*
be applied before calling XLogInsert, but heap_update and the tuple
moving routines in VACUUM FULL were ignoring this rule. A crash and
replay could therefore leave the flag incorrectly set, causing rows
to appear visible in seqscans when they should not be. This might explain
recent reports of data corruption from Jeff Ross and others.
In passing, do a bit of editorialization on comments in visibilitymap.c.
physical conversion when there are dropped columns in the same places in
the input and output tupdescs. This avoids possible performance loss from
the recent patch to improve dropped-column handling, in some cases where
the old code would have worked.
To make this work in the base case, pg_database now has a nailed-in-cache
relation descriptor that is initialized using hardwired knowledge in
relcache.c. This means pg_database is added to the set of relations that
need to have a Schema_pg_xxx macro maintained in pg_attribute.h. When this
path is taken, we'll have to do a seqscan of pg_database to find the row
we need.
In the normal case, we are able to do an indexscan to find the database's row
by name. This is made possible by storing a global relcache init file that
describes only the shared catalogs and their indexes (and therefore is usable
by all backends in any database). A new backend loads this cache file,
finds its database OID after an indexscan on pg_database, and then loads
the local relcache init file for that database.
This change should effectively eliminate number of databases as a factor
in backend startup time, even with large numbers of databases. However,
the real reason for doing it is as a first step towards getting rid of
the flat files altogether. There are still several other sub-projects
to be tackled before that can happen.
by supporting conversions in places that used to demand exact rowtype match.
Since this issue is certain to come up elsewhere (in fact, already has,
in ExecEvalConvertRowtype), factor out the support code into new core
functions for tuple conversion. I chose to put these in a new source
file since heaptuple.c is already overly long.
Heavily revised version of a patch by Pavel Stehule.
values being complained of.
In passing, also remove the arbitrary length limitation in the similar
error detail message for foreign key violations.
Itagaki Takahiro
The current implementation fires an AFTER ROW trigger for each tuple that
looks like it might be non-unique according to the index contents at the
time of insertion. This works well as long as there aren't many conflicts,
but won't scale to massive unique-key reassignments. Improving that case
is a TODO item.
Dean Rasheed
not forced out-of-line unless that is necessary to make the row fit on a
page. Previously, they were forced out-of-line if needed to get the row
down to the default target size (1/4th page).
Kevin Grittner
This alters various incidental uses of C++ key words to use other similar
identifiers, so that a C++ compiler won't choke outright. You still
(probably) need extern "C" { }; around the inclusion of backend headers.
based on a patch by Kurt Harriman <harriman@acm.org>
Also add a script cpluspluscheck to check for C++ compatibility in the
future. As of right now, this passes without error for me.
archive recovery. Invent a separate state variable and inquiry function
for XLogInsertAllowed() to clarify some tests and make the management of
writing the end-of-recovery checkpoint less klugy. Fix several places
that were incorrectly testing InRecovery when they should be looking at
RecoveryInProgress or XLogInsertAllowed (because they will now be executed
in the bgwriter not startup process). Clarify handling of bad LSNs passed
to XLogFlush during recovery. Use a spinlock for setting/testing
SharedRecoveryInProgress. Improve quite a lot of comments.
Heikki and Tom
during it:
When bgwriter is active, the startup process can't perform mdsync() correctly
because it won't see the fsync requests accumulated in bgwriter's private
pendingOpsTable. Therefore make bgwriter responsible for the end-of-recovery
checkpoint as well, when it's active.
When bgwriter is active (= archive recovery), the startup process must not
accumulate fsync requests to its own pendingOpsTable, since bgwriter won't
see them there when it performs restartpoints. Make startup process drop its
pendingOpsTable when bgwriter is launched to avoid that.
Update minimum recovery point one last time when leaving archive recovery.
It won't be updated by the end-of-recovery checkpoint because XLogFlush()
sees us as out of recovery already.
This fixes bug #4879 reported by Fujii Masao.
acting like runs inside WAL recovery, but it doesn't. I must've copy-pasted
this from a redo-function in the relation forks patch. Noticed by Tom Lane
while he was looking through callers of smgrdounlink().
node starts from the same place as the first scan did. This avoids surprising
behavior of scrollable and WITH HOLD cursors, as seen in Mark Kirkwood's bug
report of yesterday.
It's not entirely clear whether a rescan should be forced to drop out of the
syncscan mode, but for the moment I left the code behaving the same on that
point. Any change there would only be a performance and not a correctness
issue, anyway.
Back-patch to 8.3, since the unstable behavior was created by the syncscan
patch.
behavior in cases where we don't know the heap tuple count accurately; in
particular partial vacuum, but this also makes the API a bit more useful
for ANALYZE. This patch adds "estimated_count" flags to both structs so
that an approximate count can be flagged as such, and adjusts the logic
so that approximate counts are not used for updating pg_class.reltuples.
This fixes my previous complaint that VACUUM was putting ridiculous values
into pg_class.reltuples for indexes. The actual impact of that bug is
limited, because the planner only pays attention to reltuples for an index
if the index is partial; which probably explains why beta testers hadn't
noticed a degradation in plan quality from it. But it needs to be fixed.
The whole thing is a bit messy and should be redesigned in future, because
reltuples now has the potential to drift quite far away from reality when
a long period elapses with no non-partial vacuums. But this is as good as
it's going to get for 8.4.
is supposed to remove duplicate heap TIDs, we have to be sure to reduce the
tuple size and posting-item count accordingly in addItemPointersToTuple().
Failing to do so resulted in the effective injection of garbage TIDs into the
index contents, ie, whatever happened to be in the memory palloc'd for the
new tuple. I'm not sure that this fully explains the index corruption
reported by Tatsuo Ishii, but the test case I'm using no longer fails.
symbolic links with the -l option, and as Fujii Masao pointed out we ended up
overwriting files in the archive directory before this patch. Patch by
Aidan Van Dyk, Fujii Masao and me.
Backpatch to 8.3, where pg_standby was introduced.
copies?) to ensure they really don't run proc_exit/shmem_exit callbacks,
as was intended. I broke this behavior recently by installing atexit
callbacks without thinking about the one case where we truly don't want
to run those callback functions. Noted in an example from Dave Page.
is run at the end of archive recovery, providing a chance to do external
cleanup. Modify pg_standby so that it no longer removes the trigger file,
that is to be done using the recovery_end_command now.
Provide a "smart" failover mode in pg_standby, where we don't fail over
immediately, but only after recovering all unapplied WAL from the archive.
That gives you zero data loss assuming all WAL was archived before
failover, which is what most users of pg_standby actually want.
recovery_end_command by Simon Riggs, pg_standby changes by Fujii Masao and
myself.
redirecting libxml's allocations into a Postgres context. Instead, just let
it use malloc directly, and add PG_TRY blocks as needed to be sure we release
libxml data structures in error recovery code paths. This is ugly but seems
much more likely to play nicely with third-party uses of libxml, as seen in
recent trouble reports about using Perl XML facilities in pl/perl and bug
#4774 about contrib/xml2.
I left the code for allocation redirection in place, but it's only
built/used if you #define USE_LIBXMLCONTEXT. This is because I found it
useful to corral libxml's allocations in a palloc context when hunting
for libxml memory leaks, and we're surely going to have more of those
in the future with this type of approach. But we don't want it turned on
in a normal build because it breaks exactly what we need to fix.
I have not re-indented most of the code sections that are now wrapped
by PG_TRY(); that's for ease of review. pg_indent will fix it.
This is a pre-existing bug in 8.3, but I don't dare back-patch this change
until it's gotten a reasonable amount of field testing.
errors when tables are concurrently dropped. To do this we must take lock
on each relation before we check its privileges. The old code was trying
to do that the other way around, which is a bit pointless when there are lots
of other commands that lock relations before checking privileges. I did keep
it checking each relation's privilege before locking the next relation, which
is a detail that ALTER TABLE isn't too picky about.
you can end up with an unrecoverable backup if you start a new base backup
right after finishing archive recovery. In that scenario, the redo pointer of
the checkpoint that pg_start_backup() writes points to the XLOG segment where
the timeline-changing end-of-archive-recovery checkpoint is. The beginning
of that segment contains pages with the old timeline ID, and we don't accept
that in recovery unless we find a history file covering the old timeline ID.
If you omit pg_xlog from the base backup and clear the archive directory
before starting the backup, there will be no such history file available.
The bug is present in all versions since PITR was introduced in 8.0, but I'm
back-patching only back to 8.2. Earlier versions didn't have XLOG switch
records, making this fix unfeasible. Given the lack of reports until now,
it doesn't seem worthwhile to spend more effort to fix 8.0 and 8.1.
Per report and suggestion by Mikael Krantz
points where we step right or left to the next page. This should ensure
reasonable response time to a query cancel request during an unsuccessful
index scan, as seen in recent gripe from Marc Cousin. It's a bit trickier
than it might seem at first glance, because CHECK_FOR_INTERRUPTS() is a no-op
if executed while holding a buffer lock. So we have to do it just at the
point where we've dropped one page lock and not yet acquired the next.
Remove CHECK_FOR_INTERRUPTS calls at the top level of btgetbitmap and
hashgetbitmap, since they're pointless given the added checks.
I think that GIST is okay already --- at least, there's a CHECK_FOR_INTERRUPTS
at a plausible-looking place in gistnext(). I don't claim to know GIN well
enough to try to poke it for this, if indeed it has a problem at all.
This is a pre-existing issue, but in view of the lack of prior complaints
I'm not going to risk back-patching.
documentation warnings against setting it nonzero unless active use of
prepared transactions is intended and a suitable transaction manager has been
installed. This should help to prevent the type of scenario we've seen
several times now where a prepared transaction is forgotten and eventually
causes severe maintenance problems (or even anti-wraparound shutdown).
The only real reason we had the default be nonzero in the first place was to
support regression testing of the feature. To still be able to do that,
tweak pg_regress to force a nonzero value during "make check". Since we
cannot force a nonzero value in "make installcheck", add a variant regression
test "expected" file that shows the results that will be obtained when
max_prepared_transactions is zero.
Also, extend the HINT messages for transaction wraparound warnings to mention
the possibility that old prepared transactions are causing the problem.
All per today's discussion.
the checkpoint in immediate or lazy mode. This is to address complaints
that pg_start_backup() takes a long time even when there's no need to minimize
its I/O consumption.
of adding optional namespace and action fields to DefElem. Having three
node types that do essentially the same thing bloats the code and leads
to errors of confusion, such as in yesterday's bug report from Khee Chin.
To implement this without almost duplicating the reloption table, treat
relopt_kind as a bitmask instead of an integer value. This decreases the
range of allowed values, but it's not clear that there's need for that much
values anyway.
This patch also makes heap_reloptions explicitly a no-op for relation kinds
other than heap and TOAST tables.
Patch by ITAGAKI Takahiro with minor edits from me. (In particular I removed
the bit about adding relation kind to an error message, which I intend to
commit separately.)
TupleTableSlots. We have functions for retrieving a minimal tuple from a slot
after storing a regular tuple in it, or vice versa; but these were implemented
by converting the internal storage from one format to the other. The problem
with that is it invalidates any pass-by-reference Datums that were already
fetched from the slot, since they'll be pointing into the just-freed version
of the tuple. The known problem cases involve fetching both a whole-row
variable and a pass-by-reference value from a slot that is fed from a
tuplestore or tuplesort object. The added regression tests illustrate some
simple cases, but there may be other failure scenarios traceable to the same
bug. Note that the added tests probably only fail on unpatched code if it's
built with --enable-cassert; otherwise the bug leads to fetching from freed
memory, which will not have been overwritten without additional conditions.
Fix by allowing a slot to contain both formats simultaneously; which turns out
not to complicate the logic much at all, if anything it seems less contorted
than before.
Back-patch to 8.2, where minimal tuples were introduced.
method to pass extra data to the consistent() and comparePartial() methods.
This is the core infrastructure needed to support the soon-to-appear
contrib/btree_gin module. The APIs are still upward compatible with the
definitions used in 8.3 and before, although *not* with the previous 8.4devel
function definitions.
catversion bump for changes in pg_proc entries (although these are just
cosmetic, since GIN doesn't actually look at the function signature before
calling it...)
Teodor Sigaev and Oleg Bartunov
them from degrading badly when the input is sorted or nearly so. In this
scenario the tree is unbalanced to the point of becoming a mere linked list,
so insertions become O(N^2). The easiest and most safely back-patchable
solution is to stop growing the tree sooner, ie limit the growth of N. We
might later consider a rebalancing tree algorithm, but it's not clear that
the benefit would be worth the cost and complexity. Per report from Sergey
Burladyan and an earlier complaint from Heikki.
Back-patch to 8.2; older versions didn't have GIN indexes.
multiple index entries in a holding area before adding them to the main index
structure. This helps because bulk insert is (usually) significantly faster
than retail insert for GIN.
This patch also removes GIN support for amgettuple-style index scans. The
API defined for amgettuple is difficult to support with fastupdate, and
the previously committed partial-match feature didn't really work with
it either. We might eventually figure a way to put back amgettuple
support, but it won't happen for 8.4.
catversion bumped because of change in GIN's pg_am entry, and because
the format of GIN indexes changed on-disk (there's a metapage now,
and possibly a pending list).
Teodor Sigaev
some bufmgr probes, take out redundant and memory-leak-inducing path arguments
to smgr__md__read__done and smgr__md__write__done, fix bogus attempt to
recalculate space used in sort__done, clean up formatting in places where
I'm not sure pgindent will do a nice job by itself.
to 100ms (from 1000). This still seems to be comfortably larger than the
useful range of the parameter, and it should help discourage people from
picking uselessly large values. Tweak the documentation to recommend small
values, too. Per discussion of a couple weeks ago.
its usual buffer cleaning duties during archive recovery, and it's responsible
for performing restartpoints.
This requires some changes in postmaster. When the startup process has done
all the initialization and is ready to start WAL redo, it signals the
postmaster to launch the background writer. The postmaster is signaled again
when the point in recovery is reached where we know that the database is in
consistent state. Postmaster isn't interested in that at the moment, but
that's the point where we could let other backends in to perform read-only
queries. The postmaster is signaled third time when the recovery has ended,
so that postmaster knows that it's safe to start accepting connections.
The startup process now traps SIGTERM, and performs a "clean" shutdown. If
you do a fast shutdown during recovery, a shutdown restartpoint is performed,
like a shutdown checkpoint, and postmaster kills the processes cleanly. You
still have to continue the recovery at next startup, though.
Currently, the background writer is only launched during archive recovery.
We could launch it during crash recovery as well, but it seems better to keep
that codepath as simple as possible, for the sake of robustness. And it
couldn't do any restartpoints during crash recovery anyway, so it wouldn't be
that useful.
log_restartpoints is gone. Use log_checkpoints instead. This is yet to be
documented.
This whole operation is a pre-requisite for Hot Standby, but has some value of
its own whether the hot standby patch makes 8.4 or not.
Simon Riggs, with lots of modifications by me.
per-table overrides of parameters.
This removes a whole class of problems related to misusing the catalog,
and perhaps more importantly, gives us pg_dump support for the parameters.
Based on a patch by Euler Taveira de Oliveira, heavily reworked by me.
qualifier, and add support for this in pg_dump.
This allows TOAST tables to have user-defined fillfactor, and will also
enable us to move the autovacuum parameters to reloptions without taking
away the possibility of setting values for TOAST tables.
refactor the relcache code that used to do that. This allows other callers
(particularly autovacuum) to do the same without necessarily having to open
and lock a table.
be used instead of the normal exclusive lock, and make WAL redo functions
responsible for calling RestoreBkpBlocks(). They know better what kind of a
lock they need.
At the moment, this just moves things around with no functional change, but
makes the hot standby patch that's under review cleaner.
fillRelOptions routine that stores the parsed values in the struct using a
table-based approach. Per Tom suggestion. Also remove the "continue"
in HANDLE_*_RELOPTION macros, which were useless and in spirit they were
assuming too much of how the macros were going to be used. (Note that these
macros are now unused, but the intention is to introduce some usage in a
future autovacuum patch, which is why they weren't completely removed.)
Also, do not call the string validation routine when not validating. It seems
less error-prone this way, per commentary on the amoptions SGML docs.
we can get some buildfarm feedback about whether that function is still
problematic. (Note that the planned async-preread patch will not really
prove anything one way or the other in buildfarm testing, since it will
be inactive with default GUC settings.)
bitmap. This is extracted from Greg Stark's posix_fadvise patch; it seems
worth committing separately, since it's potentially useful independently of
posix_fadvise.
a more complete framework for writing custom option processing routines
by user-defined access methods.
Catalog version bumped due to the general API changes, which are going to
affect user-defined "amoptions" routines.
includes a few new ones.
- Fixed compilation errors on OS X for probes that use typedefs
- Fixed a number of probes to pass ForkNumber per the relation forks
patch
- The new probes are those that were taken out from the previous
submitted patch and required simple fixes. Will submit the other probes
that may require more discussion in a separate patch.
Robert Lor
the other major heapam.c functions. The only known consequence of this
omission is that UPDATE RETURNING failed to return the correct value for
"tableoid", as per report from KaiGai Kohei.
Back-patch to 8.2. Arguably it's wrong all the way back; but without
evidence of visible breakage before RETURNING was added, I'll desist from
patching the older branches.
vacuuming (it's not), say "database-wide VACUUM" instead of "full-database
VACUUM" in the relevant hint messages. Also, document the permissions needed
to do this. Per today's discussion.
replication patch needs a signal, but we've already used SIGUSR1 and
SIGUSR2 in normal backends. This patch allows reusing SIGUSR1 for that,
and for other purposes too if the need arises.
non-writable large objects need to have their snapshots registered on the
transaction resowner, not the current portal's, because it must persist until
the large object is closed (which the portal does not). Also, ensure that the
serializable snapshot is recorded by the transaction resource owner too, even
when a subtransaction has changed the current resource owner before
serializable is taken.
Per bug reports from Pavan Deolasee.
heap page, where a set bit indicates that all tuples on the page are
visible to all transactions, and the page therefore doesn't need
vacuuming. It is stored in a new relation fork.
Lazy vacuum uses the visibility map to skip pages that don't need
vacuuming. Vacuum is also responsible for setting the bits in the map.
In the future, this can hopefully be used to implement index-only-scans,
but we can't currently guarantee that the visibility map is always 100%
up-to-date.
In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on
each heap page, also indicating that all tuples on the page are visible to
all transactions. It's important that this flag is kept up-to-date. It
is also used to skip visibility tests in sequential scans, which gives a
small performance gain on seqscans.
that a Portal is a useful and sufficient additional argument for
CreateDestReceiver --- it just isn't, in most cases. Instead formalize
the approach of passing any needed parameters to the receiver separately.
One unexpected benefit of this change is that we can declare typedef Portal
in a less surprising location.
This patch is just code rearrangement and doesn't change any functionality.
I'll tackle the HOLD-cursor-vs-toast problem in a follow-on patch.
truncations in FSM code, call FreeSpaceMapTruncateRel from smgr_redo. To
make that cleaner from modularity point of view, move the WAL-logging one
level up to RelationTruncate, and move RelationTruncate and all the
related WAL-logging to new src/backend/catalog/storage.c file. Introduce
new RelationCreateStorage and RelationDropStorage functions that are used
instead of calling smgrcreate/smgrscheduleunlink directly. Move the
pending rel deletion stuff from smgrcreate/smgrscheduleunlink to the new
functions. This leaves smgr.c as a thin wrapper around md.c; all the
transactional stuff is now in storage.c.
This will make it easier to add new forks with similar truncation logic,
like the visibility map.
heap_form_tuple. Since this removes the last remaining caller of
heap_addheader, remove it.
Extracted from the column privileges patch from Stephen Frost, with further
code cleanups by me.
for inserting tuples in increasing TID order. It's not clear whether this
fully explains Ivan Sergio Borgonovo's complaint, but simple testing
confirms that a scan that doesn't start at block 0 can slow GIN build by
a factor of three or four.
Backpatch to 8.3. Sync scan didn't exist before that.
"base/11517/3767_fsm", instead of symbolic names like "1663/11517/3767/1",
per Alvaro's suggestion. I didn't change the messages in the higher-level
index, heap and FSM routines, though, where the fork is implicit.
(but not locked, as that would risk deadlocks). Also, make it work in a small
ring of buffers to avoid having bulk inserts trash the whole buffer arena.
Robert Haas, after an idea of Simon Riggs'.
from COMMITTED to SUBCOMMITTED during recovery. This wasn't previously
possible, but it is now due to the recent changes on clog commit protocol for
subtransactions.
Simon Riggs
allowed different processes to have different addresses for the shmem segment
in quite a long time, but there were still a few places left that used the
old coding convention. Clean them up to reduce confusion and improve the
compiler's ability to detect pointer type mismatches.
Kris Jurka
and heap_deformtuple in favor of the newer functions heap_form_tuple et al
(which do the same things but use bool control flags instead of arbitrary
char values). Eliminate the former duplicate coding of these functions,
reducing the deprecated functions to mere wrappers around the newer ones.
We can't get rid of them entirely because add-on modules probably still
contain many instances of the old coding style.
Kris Jurka
on non-full-page-image WAL records, and quite arbitrarily, only if there's
less than 20% free space on the page after the insert/update (not on HOT
updates, though). The 20% cutoff should avoid most of the overhead, when
replaying a bulk insertion, for example, while ensuring that pages that
are full are marked as full in the FSM.
This is mostly to avoid the nasty worst case scenario, where you replay
from a PITR archive, and the FSM information in the base backup is really
out of date. If there was a lot of pages that the outdated FSM claims to
have free space, but don't actually have any, the first unlucky inserter
after the recovery would traverse through all those pages, just to find
out that they're full. We didn't have this problem with the old FSM
implementation, because we simply threw the FSM information away on a
non-clean shutdown.
functions into one ReadBufferExtended function, that takes the strategy
and mode as argument. There's three modes, RBM_NORMAL which is the default
used by plain ReadBuffer(), RBM_ZERO, which replaces ZeroOrReadBuffer, and
a new mode RBM_ZERO_ON_ERROR, which allows callers to read corrupt pages
without throwing an error. The FSM needs the new mode to recover from
corrupt pages, which could happend if we crash after extending an FSM file,
and the new page is "torn".
Add fork number to some error messages in bufmgr.c, that still lacked it.
correctly set. As result, killtuple() marks as dead
wrong tuple on page. Bug was introduced by me while fixing
possible duplicates during GiST index scan.
This patch eliminates the marking of subtransactions as SUBCOMMITTED in pg_clog
during their commit; instead they remain in-progress until main transaction
commit. At main transaction commit, the commit protocol is atomic-by-page
instead of one transaction at a time. To avoid a race condition with some
subtransactions appearing committed before others in the case where they span
more than one pg_clog page, we conserve the logic that marks them subcommitted
before marking the parent committed.
Simon Riggs with minor help from me
Heikki Linnakangas
Simon Riggs
Robert Haas
Bruce Momjian
Peter Eisentraut
Tom Lane
Itagaki Takahiro
Teodor Sigaev
Andrew Dunstan
Alvaro Herrera