This was not changed in HEAD, but will be done later as part of a
pgindent run. Future pgindent runs will also do this.
Report by Tom Lane
Backpatch through all supported branches, but not HEAD
The code segment that removes the old symlink (if present) wasn't clued
into the fact that on Windows, symlinks are junction points which have
to be removed with rmdir().
Backpatch to 9.0, where the failing code was introduced.
MauMau, reviewed by Muhammad Asif Naeem and Amit Kapila
Use TransactionIdIsInProgress, then TransactionIdDidCommit, to distinguish
whether a NOTIFY message's originating transaction is in progress,
committed, or aborted. The previous coding could accept a message from a
transaction that was still in-progress according to the PGPROC array;
if the client were fast enough at starting a new transaction, it might fail
to see table rows added/updated by the message-sending transaction. Which
of course would usually be the point of receiving the message. We noted
this type of race condition long ago in tqual.c, but async.c overlooked it.
The race condition probably cannot occur unless there are multiple NOTIFY
senders in action, since an individual backend doesn't send NOTIFY signals
until well after it's done committing. But if two senders commit in close
succession, it's certainly possible that we could see the second sender's
message within the race condition window while responding to the signal
from the first one.
Per bug #9557 from Marko Tiikkaja. This patch is slightly more invasive
than what he proposed, since it removes the now-redundant
TransactionIdDidAbort call.
Back-patch to 9.0, where the current NOTIFY implementation was introduced.
If the name lookups come to different conclusions due to concurrent
activity, we might perform some parts of the DDL on a different table
than other parts. At least in the case of CREATE INDEX, this can be
used to cause the permissions checks to be performed against a
different table than the index creation, allowing for a privilege
escalation attack.
This changes the calling convention for DefineIndex, CreateTrigger,
transformIndexStmt, transformAlterTableStmt, CheckIndexCompatible
(in 9.2 and newer), and AlterTable (in 9.1 and older). In addition,
CheckRelationOwnership is removed in 9.2 and newer and the calling
convention is changed in older branches. A field has also been added
to the Constraint node (FkConstraint in 8.4). Third-party code calling
these functions or using the Constraint node will require updating.
Report by Andres Freund. Patch by Robert Haas and Andres Freund,
reviewed by Tom Lane.
Security: CVE-2014-0062
The primary role of PL validators is to be called implicitly during
CREATE FUNCTION, but they are also normal functions that a user can call
explicitly. Add a permissions check to each validator to ensure that a
user cannot use explicit validator calls to achieve things he could not
otherwise achieve. Back-patch to 8.4 (all supported versions).
Non-core procedural language extensions ought to make the same two-line
change to their own validators.
Andres Freund, reviewed by Tom Lane and Noah Misch.
Security: CVE-2014-0061
Granting a role without ADMIN OPTION is supposed to prevent the grantee
from adding or removing members from the granted role. Issuing SET ROLE
before the GRANT bypassed that, because the role itself had an implicit
right to add or remove members. Plug that hole by recognizing that
implicit right only when the session user matches the current role.
Additionally, do not recognize it during a security-restricted operation
or during execution of a SECURITY DEFINER function. The restriction on
SECURITY DEFINER is not security-critical. However, it seems best for a
user testing his own SECURITY DEFINER function to see the same behavior
others will see. Back-patch to 8.4 (all supported versions).
The SQL standards do not conflate roles and users as PostgreSQL does;
only SQL roles have members, and only SQL users initiate sessions. An
application using PostgreSQL users and roles as SQL users and roles will
never attempt to grant membership in the role that is the session user,
so the implicit right to add or remove members will never arise.
The security impact was mostly that a role member could revoke access
from others, contrary to the wishes of his own grantor. Unapproved role
member additions are less notable, because the member can still largely
achieve that by creating a view or a SECURITY DEFINER function.
Reviewed by Andres Freund and Tom Lane. Reported, independently, by
Jonas Sundman and Noah Misch.
Security: CVE-2014-0060
Various places were supposing that errno could be expected to hold still
within an ereport() nest or similar contexts. This isn't true necessarily,
though in some cases it accidentally failed to fail depending on how the
compiler chanced to order the subexpressions. This class of thinko
explains recent reports of odd failures on clang-built versions, typically
missing or inappropriate HINT fields in messages.
Problem identified by Christian Kruse, who also submitted the patch this
commit is based on. (I fixed a few issues in his patch and found a couple
of additional places with the same disease.)
Back-patch as appropriate to all supported branches.
We've always allowed CREATE TABLE to create tables in the database's default
tablespace without checking for CREATE permissions on that tablespace.
Unfortunately, the original implementation of ALTER TABLE ... SET TABLESPACE
didn't pick up on that exception.
This changes ALTER TABLE ... SET TABLESPACE to allow the database's default
tablespace without checking for CREATE rights on that tablespace, just as
CREATE TABLE works today. Users could always do this through a series of
commands (CREATE TABLE ... AS SELECT * FROM ...; DROP TABLE ...; etc), so
let's fix the oversight in SET TABLESPACE's original implementation.
The standard typanalyze functions skip over values whose detoasted size
exceeds WIDTH_THRESHOLD (1024 bytes), so as to limit memory bloat during
ANALYZE. However, we (I think I, actually :-() failed to consider the
possibility that *every* non-null value in a column is too wide. While
compute_minimal_stats() seems to behave reasonably anyway in such a case,
compute_scalar_stats() just fell through and generated no pg_statistic
entry at all. That's unnecessarily pessimistic: we can still produce
valid stanullfrac and stawidth values in such cases, since we do include
too-wide values in the average-width calculation. Furthermore, since the
general assumption in this code is that too-wide values are probably all
distinct from each other, it seems reasonable to set stadistinct to -1
("all distinct").
Per complaint from Kadri Raudsepp. This has been like this since roughly
neolithic times, so back-patch to all supported branches.
This prevents a possible longjmp out of the signal handler if a timeout
or SIGINT occurs while something within the handler has transiently set
ImmediateInterruptOK. For safety we must hold off the timeout or cancel
error until we're back in mainline, or at least till we reach the end of
the signal handler when ImmediateInterruptOK was true at entry. This
syncs these functions with the logic now present in handle_sig_alarm.
AFAICT there is no live bug here in 9.0 and up, because I don't think we
currently can wait for any heavyweight lock inside these functions, and
there is no other code (except read-from-client) that will turn on
ImmediateInterruptOK. However, that was not true pre-9.0: in older
branches ProcessIncomingNotify might block trying to lock pg_listener, and
then a SIGINT could lead to undesirable control flow. It might be all
right anyway given the relatively narrow code ranges in which NOTIFY
interrupts are enabled, but for safety's sake I'm back-patching this.
Vacuum recognizes that it can update relfrozenxid by checking whether it has
processed all pages of a relation. Unfortunately it performed that check
after truncating the dead pages at the end of the relation, and used the new
number of pages to decide whether all pages have been scanned. If the new
number of pages happened to be smaller or equal to the number of pages
scanned, it incorrectly decided that all pages were scanned.
This can lead to relfrozenxid being updated, even though some pages were
skipped that still contain old XIDs. That can lead to data loss due to xid
wraparounds with some rows suddenly missing. This likely has escaped notice
so far because it takes a large number (~2^31) of xids being used to see the
effect, while a full-table vacuum before that would fix the issue.
The incorrect logic was introduced by commit
b4b6923e03. Backpatch this fix down to 8.4,
like that commit.
Andres Freund, with some modifications by me.
Formerly, when using a SQL-spec timezone setting with a fixed GMT offset
(called a "brute force" timezone in the code), the session_timezone
variable was not updated to match the nominal timezone; rather, all code
was expected to ignore session_timezone if HasCTZSet was true. This is
of course obviously fragile, though a search of the code finds only
timeofday() failing to honor the rule. A bigger problem was that
DetermineTimeZoneOffset() supposed that if its pg_tz parameter was
pointer-equal to session_timezone, then HasCTZSet should override the
parameter. This would cause datetime input containing an explicit zone
name to be treated as referencing the brute-force zone instead, if the
zone name happened to match the session timezone that had prevailed
before installing the brute-force zone setting (as reported in bug #8572).
The same malady could affect AT TIME ZONE operators.
To fix, set up session_timezone so that it matches the brute-force zone
specification, which we can do using the POSIX timezone definition syntax
"<abbrev>offset", and get rid of the bogus lookaside check in
DetermineTimeZoneOffset(). Aside from fixing the erroneous behavior in
datetime parsing and AT TIME ZONE, this will cause the timeofday() function
to print its result in the user-requested time zone rather than some
previously-set zone. It might also affect results in third-party
extensions, if there are any that make use of session_timezone without
considering HasCTZSet, but in all cases the new behavior should be saner
than before.
Back-patch to all supported branches.
There is a rare race condition, when a transaction that inserted a tuple
aborts while vacuum is processing the page containing the inserted tuple.
Vacuum prunes the page first, which normally removes any dead tuples, but
if the inserting transaction aborts right after that, the loop after
pruning will see a dead tuple and remove it instead. That's OK, but if the
page is on a table with no indexes, and the page becomes completely empty
after removing the dead tuple (or tuples) on it, it will be immediately
marked as all-visible. That's OK, but the sanity check in vacuum would
throw a warning because it thinks that the page contains dead tuples and
was nevertheless marked as all-visible, even though it just vacuumed away
the dead tuples and so it doesn't actually contain any.
Spotted this while reading the code. It's difficult to hit the race
condition otherwise, but can be done by putting a breakpoint after the
heap_page_prune() call.
Backpatch all the way to 8.4, where this code first appeared.
Refactoring as part of commit 8ceb245680
had the unintended effect of making REINDEX TABLE and REINDEX DATABASE
no longer validate constraints enforced by the indexes in question;
REINDEX INDEX still did so. Indexes marked invalid remained so, and
constraint violations arising from data corruption went undetected.
Back-patch to 9.0, like the causative commit.
In most scenarios a portal without a ResourceOwner is dead and not subject
to any further execution, but a portal for a cursor WITH HOLD remains in
existence with no ResourceOwner after the creating transaction is over.
In this situation, if we attempt to "execute" the portal directly to fetch
data from it, we were setting CurrentResourceOwner to NULL, leading to a
segfault if the datatype output code did anything that required a resource
owner (such as trying to fetch system catalog entries that weren't already
cached). The case appears to be impossible to provoke with stock libpq,
but psqlODBC at least is able to cause it when working with held cursors.
Simplest fix is to just skip the assignment to CurrentResourceOwner, so
that any resources used by the data output operations will be managed by
the transaction-level resource owner instead. For consistency I changed
all the places that install a portal's resowner as current, even though
some of them are probably not reachable with a held cursor's portal.
Per report from Joshua Berry (with thanks to Hiroshi Inoue for developing
a self-contained test case). Back-patch to all supported versions.
Patch b19e4250b4 attempted to
preserve existing behavior regarding statistics generation in the
case that a truncation attempt was canceled due to lock conflicts.
It failed to do this accurately in two regards: (1) autovacuum had
previously generated statistics if the truncate attempt failed to
initially get the lock rather than having started the attempt, and
(2) the VACUUM ANALYZE command had always generated statistics.
Both of these changes were unintended, and are reverted by this
patch. On review, there seems to be consensus that the previous
failure to generate statistics when the truncate was terminated
was more an unfortunate consequence of how that effort was
previously terminated than a feature we want to keep; so this
patch generates statistics even when an autovacuum truncation
attempt terminates early. Another unintended change which is kept
on the basis that it is an improvement is that when a VACUUM
command is truncating, it will the new heuristic for avoiding
blocking other processes, rather than keeping an
AccessExclusiveLock on the table for however long the truncation
takes.
Per multiple reports, with some renaming per patch by Jeff Janes.
Backpatch to 9.0, where problem was created.
There was a high probability of two or more concurrent C.I.C. commands
deadlocking just before completion, because each would wait for the others
to release their reference snapshots. Fix by releasing the snapshot
before waiting for other snapshots to go away.
Per report from Paul Hinze. Back-patch to all active branches.
Since a backend adds itself to the global listener array during
Exec_ListenPreCommit, it's inappropriate for it to remove itself during
Exec_UnlistenCommit or Exec_UnlistenAllCommit --- that leads to failure
when committing a transaction that did UNLISTEN then LISTEN, since we end
up not registered though we should be. (This leads to missing later
notifications, or to Assert failures in assert-enabled builds.) Instead
deal with deregistering at the bottom of AtCommit_Notify, when we know the
final state of the listenChannels list.
Also, simplify the representation of registration status by replacing the
transient backendHasExecutedInitialListen flag with an amRegisteredListener
flag.
Per report from Greg Sabino Mullane. Back-patch to 9.0, where the problem
was introduced during the LISTEN/NOTIFY rewrite.
The original code used freeze_min_age instead of freeze_table_age. The
main consequence of this mistake is that lowering freeze_min_age would
cause full-table scans to occur much more frequently, which causes
serious issues because the number of writes required is much larger.
That feature (freeze_min_age) is supposed to affect only how soon tuples
are frozen; some pages should still be skipped due to the visibility
map.
Backpatch to 8.4, where the freeze_table_age feature was introduced.
Report and patch from Andres Freund
In situations where there are over 8MB of empty pages at the end of
a table, the truncation work for trailing empty pages takes longer
than deadlock_timeout, and there is frequent access to the table by
processes other than autovacuum, there was a problem with the
autovacuum worker process being canceled by the deadlock checking
code. The truncation work done by autovacuum up that point was
lost, and the attempt tried again by a later autovacuum worker. The
attempts could continue indefinitely without making progress,
consuming resources and blocking other processes for up to
deadlock_timeout each time.
This patch has the autovacuum worker checking whether it is
blocking any other thread at 20ms intervals. If such a condition
develops, the autovacuum worker will persist the work it has done
so far, release its lock on the table, and sleep in 50ms intervals
for up to 5 seconds, hoping to be able to re-acquire the lock and
try again. If it is unable to get the lock in that time, it moves
on and a worker will try to continue later from the point this one
left off.
While this patch doesn't change the rules about when and what to
truncate, it does cause the truncation to occur sooner, with less
blocking, and with the consumption of fewer resources when there is
contention for the table's lock.
The only user-visible change other than improved performance is
that the table size during truncation may change incrementally
instead of just once.
Backpatched to 9.0 from initial master commit at
b19e4250b4 -- before that the
differences are too large to be clearly safe.
Jan Wieck
Use of SnapshotNow is known to expose us to race conditions if the tuple(s)
being sought could be updated by concurrently-committing transactions.
CREATE DATABASE and DROP DATABASE are particularly exposed because they do
heavyweight filesystem operations during their scans of pg_tablespace,
so that the scans run for a very long time compared to most. Furthermore,
the potential consequences of a missed or twice-visited row are nastier
than average:
* createdb() could fail with a bogus "file already exists" error, or
silently fail to copy one or more tablespace's worth of files into the
new database.
* remove_dbtablespaces() could miss one or more tablespaces, thus failing
to free filesystem space for the dropped database.
* check_db_file_conflict() could likewise miss a tablespace, leading to an
OID conflict that could result in data loss either immediately or in
future operations. (This seems of very low probability, though, since a
duplicate database OID would be unlikely to start with.)
Hence, it seems worth fixing these three places to use MVCC snapshots, even
though this will someday be superseded by a generic solution to SnapshotNow
race conditions.
Back-patch to all active branches.
Stephen Frost and Tom Lane
During VACUUM if we pause to perform a cycle
of index cleanup we drop the vmbuffer pin,
so we should do the same thing when heap
scan completes. This avoids holding vmbuffer
pin across the main index cleanup in VACUUM,
which could be minutes or hours longer than
necessary for correctness.
Bug report and suggested fix from Pavan Deolasee
If we had not been holding buffer pin continuously since the tuple was
initially fetched by the UPDATE or DELETE query, it would be possible for
VACUUM or a page-prune operation to move the tuple while we're trying to
copy it. This would result in a garbage "old" tuple value being passed to
an AFTER ROW UPDATE or AFTER ROW DELETE trigger. The preconditions for
this are somewhat improbable, and the timing constraints are very tight;
so it's not so surprising that this hasn't been reported from the field,
even though the bug has been there a long time.
Problem found by Andres Freund. Back-patch to all active branches.
This patch changes CREATE INDEX CONCURRENTLY so that the pg_index
flag changes it makes without exclusive lock on the index are made via
heap_inplace_update() rather than a normal transactional update. The
latter is not very safe because moving the pg_index tuple could result in
concurrent SnapshotNow scans finding it twice or not at all, thus possibly
resulting in index corruption.
In addition, fix various places in the code that ought to check to make
sure that the indexes they are manipulating are valid and/or ready as
appropriate. These represent bugs that have existed since 8.2, since
a failed CREATE INDEX CONCURRENTLY could leave a corrupt or invalid
index behind, and we ought not try to do anything that might fail with
such an index.
Also fix RelationReloadIndexInfo to ensure it copies all the pg_index
columns that are allowed to change after initial creation. Previously we
could have been left with stale values of some fields in an index relcache
entry. It's not clear whether this actually had any user-visible
consequences, but it's at least a bug waiting to happen.
This is a subset of a patch already applied in 9.2 and HEAD. Back-patch
into all earlier supported branches.
Tom Lane and Andres Freund
This case got broken in 8.4 by the addition of an error check that
complains if ALTER TABLE ONLY is used on a table that has children.
We do use ONLY for this situation, but it's okay because the necessary
recursion occurs at a higher level. So we need to have a separate
flag to suppress recursion without making the error check.
Reported and patched by Pavan Deolasee, with some editorial adjustments by
me. Back-patch to 8.4, since this is a regression of functionality that
worked in earlier branches.
If a crash occurred immediately after the first nextval() call for a serial
column, WAL replay would restore the sequence to a state in which it
appeared that no nextval() had been done, thus allowing the first sequence
value to be returned again by the next nextval() call; as reported in
bug #6748 from Xiangming Mei.
More generally, the problem would occur if an ALTER SEQUENCE was executed
on a freshly created or reset sequence. (The manifestation with serial
columns was introduced in 8.2 when we added an ALTER SEQUENCE OWNED BY step
to serial column creation.) The cause is that sequence creation attempted
to save one WAL entry by writing out a WAL record that made it appear that
the first nextval() had already happened (viz, with is_called = true),
while marking the sequence's in-database state with log_cnt = 1 to show
that the first nextval() need not emit a WAL record. However, ALTER
SEQUENCE would emit a new WAL entry reflecting the actual in-database state
(with is_called = false). Then, nextval would allocate the first sequence
value and set is_called = true, but it would trust the log_cnt value and
not emit any WAL record. A crash at this point would thus restore the
sequence to its post-ALTER state, causing the next nextval() call to return
the first sequence value again.
To fix, get rid of the idea of logging an is_called status different from
reality. This means that the first nextval-driven WAL record will happen
at the first nextval call not the second, but the marginal cost of that is
pretty negligible. In addition, make sure that ALTER SEQUENCE resets
log_cnt to zero in any case where it touches sequence parameters that
affect future nextval results. This will result in some user-visible
changes in the contents of a sequence's log_cnt column, as reflected in the
patch's regression test changes; but no application should be depending on
that anyway, since it was already true that log_cnt changes rather
unpredictably depending on checkpoint timing.
In addition, make some basically-cosmetic improvements to get rid of
sequence.c's undesirable intimacy with page layout details. It was always
really trying to WAL-log the contents of the sequence tuple, so we should
have it do that directly using a HeapTuple's t_data and t_len, rather than
backing into it with some magic assumptions about where the tuple would be
on the sequence's page.
Back-patch to all supported branches.
If a CHECK constraint or index definition contained a whole-row Var (that
is, "table.*"), an attempt to copy that definition via CREATE TABLE LIKE or
table inheritance produced incorrect results: the copied Var still claimed
to have the rowtype of the source table, rather than the created table.
For the LIKE case, it seems reasonable to just throw error for this
situation, since the point of LIKE is that the new table is not permanently
coupled to the old, so there's no reason to assume its rowtype will stay
compatible. In the inheritance case, we should ideally allow such
constraints, but doing so will require nontrivial refactoring of CREATE
TABLE processing (because we'd need to know the OID of the new table's
rowtype before we adjust inherited CHECK constraints). In view of the lack
of previous complaints, that doesn't seem worth the risk in a back-patched
bug fix, so just make it throw error for the inheritance case as well.
Along the way, replace change_varattnos_of_a_node() with a more robust
function map_variable_attnos(), which is capable of being extended to
handle insertion of ConvertRowtypeExpr whenever we get around to fixing
the inheritance case nicely, and in the meantime it returns a failure
indication to the caller so that a helpful message with some context can be
thrown. Also, this code will do the right thing with subselects (if we
ever allow them in CHECK or indexes), and it range-checks varattnos before
using them to index into the map array.
Per report from Sergey Konoplev. Back-patch to all supported branches.
The LISTEN/NOTIFY subsystem got confused if SimpleLruZeroPage failed,
which would typically happen as a result of a write() failure while
attempting to dump a dirty pg_notify page out of memory. Subsequently,
all attempts to send more NOTIFY messages would fail with messages like
"Could not read from file "pg_notify/nnnn" at offset nnnnn: Success".
Only restarting the server would clear this condition. Per reports from
Kevin Grittner and Christoph Berg.
Back-patch to 9.0, where the problem was introduced during the
LISTEN/NOTIFY rewrite.
If the tablespace directory is missing entirely, we allow DROP TABLESPACE
to go through, on the grounds that it should be possible to clean up the
catalog entry in such a situation. However, we forgot that the pg_tblspc
symlink might still be there. We should try to remove the symlink too
(but not fail if it's no longer there), since not doing so can lead to
weird behavior subsequently, as per report from Michael Nolan.
There was some discussion of adding dependency links to prevent DROP
TABLESPACE when the catalogs still contain references to the tablespace.
That might be worth doing too, but it's an orthogonal question, and in
any case wouldn't be back-patchable.
Back-patch to 9.0, which is as far back as the logic looks like this.
We could possibly do something similar in 8.x, but given the lack of
reports I'm not sure it's worth the trouble, and anyway the case could
not arise in the form the logic is meant to cover (namely, a post-DROP
transaction rollback having resurrected the pg_tablespace entry after
some or all of the filesystem infrastructure is gone).
The COPY documentation says "COPY FROM matches the input against the null
string before removing backslashes". It is therefore reasonable to presume
that null markers like E'\\0' will work ... and they did, until someone put
the tests in the wrong order during microoptimization-driven rewrites.
Since then, we've been failing if the null marker is something that would
de-escape to an invalidly-encoded string. Since null markers generally
need to be something that can't appear in the data, this represents a
nontrivial loss of functionality; surprising nobody noticed it earlier.
Per report from Jeff Davis. Backpatch to 8.4 where this got broken.
This check was overlooked when we added function execute permissions to the
system years ago. For an ordinary trigger function it's not a big deal,
since trigger functions execute with the permissions of the table owner,
so they couldn't do anything the user issuing the CREATE TRIGGER couldn't
have done anyway. However, if a trigger function is SECURITY DEFINER,
that is not the case. The lack of checking would allow another user to
install it on his own table and then invoke it with, essentially, forged
input data; which the trigger function is unlikely to realize, so it might
do something undesirable, for instance insert false entries in an audit log
table.
Reported by Dinesh Kumar, patch by Robert Haas
Security: CVE-2012-0866
Although we will not even issue an XLOG_TBLSPC_DROP WAL record unless
removal of the tablespace's directories succeeds, that does not guarantee
that the same operation will succeed during WAL replay. Foreseeable
reasons for it to fail include temp files created in the tablespace by Hot
Standby backends, wrong directory permissions on a standby server, etc etc.
The original coding threw ERROR if replay failed to remove the directories,
but that is a serious overreaction. Throwing an error aborts recovery,
and worse means that manual intervention will be needed to get the database
to start again, since otherwise the same error will recur on subsequent
attempts to replay the same WAL record. And the consequence of failing to
remove the directories is only that some probably-small amount of disk
space is wasted, so it hardly seems justified to throw an error.
Accordingly, arrange to report such failures as LOG messages and keep going
when a failure occurs during replay.
Back-patch to 9.0 where Hot Standby was introduced. In principle such
problems can occur in earlier releases, but Hot Standby increases the odds
of trouble significantly. Given the lack of field reports of such issues,
I'm satisfied with patching back as far as the patch applies easily.
RestoreBkpBlocks was in the habit of zeroing and refilling the target
buffer; which was perfectly safe when the code was written, but is unsafe
during Hot Standby operation. The reason is that we have coding rules
that allow backends to continue accessing a tuple in a heap relation while
holding only a pin on its buffer. Such a backend could see transiently
zeroed data, if WAL replay had occasion to change other data on the page.
This has been shown to be the cause of bug #6425 from Duncan Rance (who
deserves kudos for developing a sufficiently-reproducible test case) as
well as Bridget Frey's re-report of bug #6200. It most likely explains the
original report as well, though we don't yet have confirmation of that.
To fix, change the code so that only bytes that are supposed to change will
change, even transiently. This actually saves cycles in RestoreBkpBlocks,
since it's not writing the same bytes twice.
Also fix seq_redo, which has the same disease, though it has to work a bit
harder to meet the requirement.
So far as I can tell, no other WAL replay routines have this type of bug.
In particular, the index-related replay routines, which would certainly be
broken if they had to meet the same standard, are not at risk because we
do not have coding rules that allow access to an index page when not
holding a buffer lock on it.
Back-patch to 9.0 where Hot Standby was added.
When default_text_search_config, default_tablespace, or temp_tablespaces
setting is set per-user or per-database, with an "ALTER USER/DATABASE SET
..." statement, don't throw an error if the text search configuration or
tablespace does not exist. In case of text search configuration, even if
it doesn't exist in the current database, it might exist in another
database, where the setting is intended to have its effect. This behavior
is now the same as search_path's.
Tablespaces are cluster-wide, so the same argument doesn't hold for
tablespaces, but there's a problem with pg_dumpall: it dumps "ALTER USER
SET ..." statements before the "CREATE TABLESPACE" statements. Arguably
that's pg_dumpall's fault - it should dump the statements in such an order
that the tablespace is created first and then the "ALTER USER SET
default_tablespace ..." statements after that - but it seems better to be
consistent with search_path and default_text_search_config anyway. Besides,
you could still create a dump that throws an error, by creating the
tablespace, running "ALTER USER SET default_tablespace", then dropping the
tablespace and running pg_dumpall on that.
Backpatch to all supported versions.
In commit 7b0d0e9356, I made CLUSTER and
VACUUM FULL try to preserve toast value OIDs from the original toast table
to the new one. However, if we have to copy both live and recently-dead
versions of a row that has a toasted column, those versions may well
reference the same toast value with the same OID. The patch then led to
duplicate-key failures as we tried to insert the toast value twice with the
same OID. (The previous behavior was not very desirable either, since it
would have silently inserted the same value twice with different OIDs.
That wastes space, but what's worse is that the toast values inserted for
already-dead heap rows would not be reclaimed by subsequent ordinary
VACUUMs, since they go into the new toast table marked live not deleted.)
To fix, check if the copied OID already exists in the new toast table, and
if so, assume that it stores the desired value. This is reasonably safe
since the only case where we will copy an OID from a previous toast pointer
is when toast_insert_or_update was given that toast pointer and so we just
pulled the data from the old table; if we got two different values that way
then we have big problems anyway. We do have to assume that no other
backend is inserting items into the new toast table concurrently, but
that's surely safe for CLUSTER and VACUUM FULL.
Per bug #6393 from Maxim Boguk. Back-patch to 9.0, same as the previous
patch.
We forgot to modify column ACLs, so privileges were still shown as having
been granted by the old owner. This meant that neither the new owner nor
a superuser could revoke the now-untraceable-to-table-owner permissions.
Per bug #6350 from Marc Balmer.
This has been wrong since column ACLs were added, so back-patch to 8.4.
When a foreign-key constraint references another column of the same table,
row updates will queue both the PK's ON UPDATE action and the FK's CHECK
action in the same event. The ON UPDATE action must execute first, else
the CHECK will check a non-final state of the row and possibly throw an
inappropriate error, as seen in bug #6268 from Roman Lytovchenko.
Now, the firing order of multiple triggers for the same event is determined
by the sort order of their pg_trigger.tgnames, and the auto-generated names
we use for FK triggers are "RI_ConstraintTrigger_NNNN" where NNNN is the
trigger OID. So most of the time the firing order is the same as creation
order, and so rearranging the creation order fixes it.
This patch will fail to fix the problem if the OID counter wraps around or
adds a decimal digit (eg, from 99999 to 100000) while we are creating the
triggers for an FK constraint. Given the small odds of that, and the low
usage of self-referential FKs, we'll live with that solution in the back
branches. A better fix is to change the auto-generated names for FK
triggers, but it seems unwise to do that in stable branches because there
may be client code that depends on the naming convention. We'll fix it
that way in HEAD in a separate patch.
Back-patch to all supported branches, since this bug has existed for a long
time.
Since the last couple of columns of pg_type are often NULL,
sizeof(FormData_pg_type) can be an overestimate of the actual size of the
tuple data part. Therefore memcpy'ing that much out of the catalog cache,
as analyze.c was doing, poses a small risk of copying past the end of
memory and incurring SIGSEGV. No such crash has been identified in the
field, but we've certainly seen the equivalent happen in other code paths,
so patch this one all the way back.
Per valgrind testing by Noah Misch, though this is not his proposed patch.
I chose to use SearchSysCacheCopy1 rather than inventing special-purpose
infrastructure for copying only the minimal part of a pg_type tuple.
It is possible for VACUUM to scan no pages at all, if the visibility map
shows that all pages are all-visible. In this situation VACUUM has no new
information to report about the relation's tuple density, so it wasn't
changing pg_class.reltuples ... but it updated pg_class.relpages anyway.
That's wrong in general, since there is no evidence to justify changing the
density ratio reltuples/relpages, but it's particularly bad if the previous
state was relpages=reltuples=0, which means "unknown tuple density".
We just replaced "unknown" with "zero". ANALYZE would eventually recover
from this, but it could take a lot of repetitions of ANALYZE to do so if
the relation size is much larger than the maximum number of pages ANALYZE
will scan, because of the moving-average behavior introduced by commit
b4b6923e03.
The only known situation where we could have relpages=reltuples=0 and yet
the visibility map asserts everything's visible is immediately following
a pg_upgrade. It might be advisable for pg_upgrade to try to preserve the
relpages/reltuples statistics; but in any case this code is wrong on its
own terms, so fix it. Per report from Sergey Koposov.
Back-patch to 8.4, where the visibility map was introduced, same as the
previous change.
Due to tuple-slot mismanagement, evaluation of WHEN conditions for AFTER
ROW UPDATE triggers could crash if there had been a BEFORE ROW trigger
fired for the same update. Fix by not trying to overload the use of
estate->es_trig_tuple_slot. Per report from Yoran Heling.
Back-patch to 9.0, when trigger WHEN conditions were introduced.
This works around the problem that a catalog cache entry might contain a
toast pointer that we try to dereference just as a VACUUM FULL completes
on that catalog. We will see the sinval message on the cache entry when
we acquire lock on the toast table, but by that point we've already told
tuptoaster.c "here's the pointer to fetch", so it's difficult from a code
structural standpoint to update the pointer before we use it. Much less
painful to ensure that toast pointers are not invalidated in the first
place. We have to add a bit of code to deal with the case that a value
that previously wasn't toasted becomes so; but that should be a
seldom-exercised corner case, so the inefficiency shouldn't be significant.
Back-patch to 9.0. In prior versions, we didn't allow CLUSTER on system
catalogs, and VACUUM FULL didn't result in reassignment of toast OIDs, so
there was no problem.
When updating or deleting a system catalog tuple, it's necessary to acquire
RowExclusiveLock on the catalog before looking up the tuple; otherwise a
concurrent VACUUM FULL on the catalog might move the tuple to a different
TID before we can apply the update. Coding patterns that find the tuple
via a table scan aren't at risk here, but when obtaining the tuple from a
catalog cache, correct ordering is important; and several routines in
foreigncmds.c got it wrong. Noted while running the regression tests in
parallel with VACUUM FULL of assorted system catalogs.
For consistency I moved all the heap_open calls to the starts of their
functions, including a couple for which there was no actual bug.
Back-patch to 8.4 where foreigncmds.c was added.
It is possible for a NestLoop plan node to pass an OUTER Var into an
"inner indexscan" that is an Append construct (derived from an inheritance
tree or UNION ALL subquery). The OUTER tuple is then passed down at
runtime to the leaf indexscan node(s) where it will actually be used.
EXPLAIN has to likewise pass the information about the nestloop's outer
subplan down through the Append node, else it will fail to print the
outer-reference Vars (with complaints like "bogus varno: 65001").
However, there was a case missed in all this: we could also have gating
Result nodes that were inserted into the appendrel plan tree to deal with
pseudoconstant qual conditions. So EXPLAIN has to pass down the outer plan
node to a Result's subplan, too. Per example from Jon Nelson.
The problem is gone in 9.1 because we replaced the nestloop outer-tuple
kluge with a Param-based data transfer mechanism. Also, so far as I can
tell, the case can't happen before 8.4 because of restrictions on what
sorts of appendrel members could be pulled up into the parent query.
So this patch is only needed for 8.4 and 9.0.
I mis-simplified the test where ANALYZE decided if it could get away
without doing anything: under the new regime, that's never allowed. Per
bug #6068 from Jeff Janes. Back-patch to 8.4, just like previous patch.
My previous commit disallowed this operation, but did nothing about
cleaning up the damage if one had already been done. With the operation
disallowed, it's okay to just forcibly clear xmax in a sequence's tuple,
since any value seen there could not represent a live transaction's lock.
So, any sequence-specific operation will repair the problem automatically,
whether or not the user has already seen "could not access status of
transaction" failures.
When we added the ability for vacuum to skip heap pages by consulting the
visibility map, we made it just not update the reltuples/relpages
statistics if it skipped any pages. But this could leave us with extremely
out-of-date stats for a table that contains any unchanging areas,
especially for TOAST tables which never get processed by ANALYZE. In
particular this could result in autovacuum making poor decisions about when
to process the table, as in recent report from Florian Helmberger. And in
general it's a bad idea to not update the stats at all. Instead, use the
previous values of reltuples/relpages as an estimate of the tuple density
in unvisited pages. This approach results in a "moving average" estimate
of reltuples, which should converge to the correct value over multiple
VACUUM and ANALYZE cycles even when individual measurements aren't very
good.
This new method for updating reltuples is used by both VACUUM and ANALYZE,
with the result that we no longer need the grotty interconnections that
caused ANALYZE to not update the stats depending on what had happened
in the parent VACUUM command.
Also, fix the logic for skipping all-visible pages during VACUUM so that it
looks ahead rather than behind to decide what to do, as per a suggestion
from Greg Stark. This eliminates useless scanning of all-visible pages at
the start of the relation or just after a not-all-visible page. In
particular, the first few pages of the relation will not be invariably
included in the scanned pages, which seems to help in not overweighting
them in the reltuples estimate.
Back-patch to 8.4, where the visibility map was introduced.
This is necessary to avoid long-term memory leakage, because the main loop
in PostgresMain expects to be executing in MessageContext, and hence is a
bit sloppy about freeing stuff that is only needed for the duration of
processing the current client message. The known case of an actual leak
is when encoding conversion has to be done on the incoming command string,
but there might be others. Per report from Per-Olov Esgard.
Back-patch to 9.0, where the bug was introduced by the LISTEN/NOTIFY
rewrite.
