The previous code tried to synchronize by unlinking the init file twice,
but that doesn't actually work: it leaves a window wherein a third process
could read the already-stale init file but miss the SI messages that would
tell it the data is stale. The result would be bizarre failures in catalog
accesses, typically "could not read block 0 in file ..." later during
startup.
Instead, hold RelCacheInitLock across both the unlink and the sending of
the SI messages. This is more straightforward, and might even be a bit
faster since only one unlink call is needed.
This has been wrong since it was put in (in 2002!), so back-patch to all
supported releases.
This means that they can initially be added to a large existing table
without checking its initial contents, but new tuples must comply to
them; a separate pass invoked by ALTER TABLE / VALIDATE can verify
existing data and ensure it complies with the constraint, at which point
it is marked validated and becomes a normal part of the table ecosystem.
An non-validated CHECK constraint is ignored in the planner for
constraint_exclusion purposes; when validated, cached plans are
recomputed so that partitioning starts working right away.
This patch also enables domains to have unvalidated CHECK constraints
attached to them as well by way of ALTER DOMAIN / ADD CONSTRAINT / NOT
VALID, which can later be validated with ALTER DOMAIN / VALIDATE
CONSTRAINT.
Thanks to Thom Brown, Dean Rasheed and Jaime Casanova for the various
reviews, and Robert Hass for documentation wording improvement
suggestions.
This patch was sponsored by Enova Financial.
Opening a catcache's index could require reading from that cache's own
catalog, which of course would acquire AccessShareLock on the catalog.
So the original coding here risks locking index before heap, which could
deadlock against another backend trying to get exclusive locks in the
normal order. Because InitCatCachePhase2 is only called when a backend
has to start up without a relcache init file, the deadlock was seldom seen
in the field. (And by the same token, there's no need to worry about any
performance disadvantage; so not much point in trying to distinguish
exactly which catalogs have the risk.)
Bug report, diagnosis, and patch by Nikhil Sontakke. Additional commentary
by me. Back-patch to all supported branches.
This adds collation support for columns and domains, a COLLATE clause
to override it per expression, and B-tree index support.
Peter Eisentraut
reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
The contents of an unlogged table are WAL-logged; thus, they are not
available on standby servers and are truncated whenever the database
system enters recovery. Indexes on unlogged tables are also unlogged.
Unlogged GiST indexes are not currently supported.
This commit replaces pg_class.relistemp with pg_class.relpersistence;
and also modifies the RangeVar node type to carry relpersistence rather
than istemp. It also removes removes rd_istemp from RelationData and
instead performs the correct computation based on relpersistence.
For clarity, we add three new macros: RelationNeedsWAL(),
RelationUsesLocalBuffers(), and RelationUsesTempNamespace(), so that we
can clarify the purpose of each check that previous depended on
rd_istemp.
This is intended as infrastructure for the upcoming unlogged tables
patch, as well as for future possible work on global temporary tables.
Formerly we looked up the operators associated with each index (caching
them in relcache) and then the planner looked up the btree opfamily
containing such operators in order to build the btree-centric pathkey
representation that describes the index's sort order. This is quite
pointless for btree indexes: we might as well just use the index's opfamily
information directly. That saves syscache lookup cycles during planning,
and furthermore allows us to eliminate the relcache's caching of operators
altogether, which may help in reducing backend startup time.
I added code to plancat.c to perform the same type of double lookup
on-the-fly if it's ever faced with a non-btree amcanorder index AM.
If such a thing actually becomes interesting for production, we should
replace that logic with some more-direct method for identifying the
corresponding btree opfamily; but it's not worth spending effort on now.
There is considerably more to do pursuant to my recent proposal to get rid
of sort-operator-based representations of sort orderings, but this patch
grabs some of the low-hanging fruit. I'll look at the remainder of that
work after the current commitfest.
In the previous coding, we simply issued ALTER SEQUENCE RESTART commands,
which do not roll back on error. This meant that an error between
truncating and committing left the sequences out of sync with the table
contents, with potentially bad consequences as were noted in a Warning on
the TRUNCATE man page.
To fix, create a new storage file (relfilenode) for a sequence that is to
be reset due to RESTART IDENTITY. If the transaction aborts, we'll
automatically revert to the old storage file. This acts just like a
rewriting ALTER TABLE operation. A penalty is that we have to take
exclusive lock on the sequence, but since we've already got exclusive lock
on its owning table, that seems unlikely to be much of a problem.
The interaction of this with usual nontransactional behaviors of sequence
operations is a bit weird, but it's hard to see what would be completely
consistent. Our choice is to discard cached-but-unissued sequence values
both when the RESTART is executed, and at rollback if any; but to not touch
the currval() state either time.
In passing, move the sequence reset operations to happen before not after
any AFTER TRUNCATE triggers are fired. The previous ordering was not
logically sensible, but was forced by the need to minimize inconsistency
if the triggers caused an error. Transactional rollback is a much better
solution to that.
Patch by Steve Singer, rather heavily adjusted by me.
SI invalidation events, rather than indirectly through the relcache.
In the previous coding, we had to flush a composite-type typcache entry
whenever we discarded the corresponding relcache entry. This caused problems
at least when testing with RELCACHE_FORCE_RELEASE, as shown in recent report
from Jeff Davis, and might result in real-world problems given the kind of
unexpected relcache flush that that test mechanism is intended to model.
The new coding decouples relcache and typcache management, which is a good
thing anyway from a structural perspective. The cost is that we have to
search the typcache linearly to find entries that need to be flushed. There
are a couple of ways we could avoid that, but at the moment it's not clear
it's worth any extra trouble, because the typcache contains very few entries
in typical operation.
Back-patch to 8.2, the same as some other recent fixes in this general area.
The patch could be carried back to 8.0 with some additional work, but given
that it's only hypothetical whether we're fixing any problem observable in
the field, it doesn't seem worth the work now.
This allows us to reliably remove all leftover temporary relation
files on cluster startup without reference to system catalogs or WAL;
therefore, we no longer include temporary relations in XLOG_XACT_COMMIT
and XLOG_XACT_ABORT WAL records.
Since these changes require including a backend ID in each
SharedInvalSmgrMsg, the size of the SharedInvalidationMessage.id
field has been reduced from two bytes to one, and the maximum number
of connections has been reduced from INT_MAX / 4 to 2^23-1. It would
be possible to remove these restrictions by increasing the size of
SharedInvalidationMessage by 4 bytes, but right now that doesn't seem
like a good trade-off.
Review by Jaime Casanova and Tom Lane.
database to connect to. This is necessary for the walsender code to work
properly (it was previously using an untenable assumption that template1 would
always be available to connect to). This also gets rid of a small security
shortcoming that was introduced in the original patch to eliminate the flat
authentication files: before, you could find out whether or not the requested
database existed even if you couldn't pass the authentication checks.
The changes needed to support this are mainly just to treat pg_authid and
pg_auth_members as nailed relations, so that we can read them without having
to be able to locate real pg_class entries for them. This mechanism was
already debugged for pg_database, but we hadn't recognized the value of
applying it to those catalogs too.
Since the current code doesn't have support for accessing toast tables before
we've brought up all of the relcache, remove pg_authid's toast table to ensure
that no one can store an out-of-line toasted value of rolpassword. The case
seems quite unlikely to occur in practice, and was effectively unsupported
anyway in the old "flatfiles" implementation.
Update genbki.pl to actually implement the same rules as bootstrap.c does for
not-nullability of catalog columns. The previous coding was a bit cheesy but
worked all right for the previous set of bootstrap catalogs. It does not work
for pg_authid, where rolvaliduntil needs to be nullable.
Initdb forced due to minor catalog changes (mainly the toast table removal).
relcache reload works. In the patched code, a relcache entry in process of
being rebuilt doesn't get unhooked from the relcache hash table; which means
that if a cache flush occurs due to sinval queue overrun while we're
rebuilding it, the entry could get blown away by RelationCacheInvalidate,
resulting in crash or misbehavior. Fix by ensuring that an entry being
rebuilt has positive refcount, so it won't be seen as a target for removal
if a cache flush occurs. (This will mean that the entry gets rebuilt twice
in such a scenario, but that's okay.) It appears that the problem can only
arise within a transaction that has previously reassigned the relfilenode of
a pre-existing table, via TRUNCATE or a similar operation. Per bug #5412
from Rusty Conover.
Back-patch to 8.2, same as the patch that introduced the problem.
I think that the failure can't actually occur in 8.2, since it lacks the
rd_newRelfilenodeSubid optimization, but let's make it work like the later
branches anyway.
Patch by Heikki, slightly editorialized 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.
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.
needed by nothing else.
The restructuring I just finished doing on cache management exposed to me how
silly this routine was. Its function was to go into the catcache and blow
away all entries related to a given relation when there was a relcache flush
on that relation. However, there is no point in removing a catcache entry
if the catalog row it represents is still valid --- and if it isn't valid,
there must have been a catcache entry flush on it, because that's triggered
directly by heap_update or heap_delete on the catalog row. So this routine
accomplished nothing except to blow away valid cache entries that we'd very
likely be wanting in the near future to help reconstruct the relcache entry.
Dumb.
On top of which, it required a subtle and easy-to-get-wrong attribute in
syscache definitions, ie, the column containing the OID of the related
relation if any. Removing that is a very useful maintenance simplification.
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.
the relfilenode of currently-not-relocatable system catalogs.
1. Get rid of inval.c's dependency on relfilenode, by not having it emit
smgr invalidations as a result of relcache flushes. Instead, smgr sinval
messages are sent directly from smgr.c when an actual relation delete or
truncate is done. This makes considerably more structural sense and allows
elimination of a large number of useless smgr inval messages that were
formerly sent even in cases where nothing was changing at the
physical-relation level. Note that this reintroduces the concept of
nontransactional inval messages, but that's okay --- because the messages
are sent by smgr.c, they will be sent in Hot Standby slaves, just from a
lower logical level than before.
2. Move setNewRelfilenode out of catalog/index.c, where it never logically
belonged, into relcache.c; which is a somewhat debatable choice as well but
better than before. (I considered catalog/storage.c, but that seemed too
low level.) Rename to RelationSetNewRelfilenode.
3. Cosmetic cleanups of some other relfilenode manipulations.
underlying catalog not only the index itself. Otherwise, if the cache
load process touches the catalog (which will happen for many though not
all of these indexes), we are locking index before parent table, which can
result in a deadlock against processes that are trying to lock them in the
normal order. Per today's failure on buildfarm member gothic_moth; it's
surprising the problem hadn't been identified before.
Back-patch to 8.2. Earlier releases didn't have the issue because they
didn't try to lock these indexes during load (instead assuming that they
couldn't change schema at all during multiuser operation).
occurring during a reload, such as query-cancel. Instead of zeroing out
an existing relcache entry and rebuilding it in place, build a new relcache
entry, then swap its contents with the old one, then free the new entry.
This avoids problems with code believing that a previously obtained pointer
to a cache entry must still reference a valid entry, as seen in recent
failures on buildfarm member jaguar. (jaguar is using CLOBBER_CACHE_ALWAYS
which raises the probability of failure substantially, but the problem
could occur in the field without that.) The previous design was okay
when it was made, but subtransactions and the ResourceOwner mechanism
make it unsafe now.
Also, make more use of the already existing rd_isvalid flag, so that we
remember that the entry requires rebuilding even if the first attempt fails.
Back-patch as far as 8.2. Prior versions have enough issues around relcache
reload anyway (due to inadequate locking) that fixing this one doesn't seem
worthwhile.
can upgrade clusters without renaming the tablespace directories. New
directory structure format is, e.g.:
$PGDATA/pg_tblspc/20981/PG_8.5_201001061/719849/83292814
pg_attribute, by having genbki.pl derive the information from the various
catalog header files. This greatly simplifies modification of the
"bootstrapped" catalogs.
This patch finally kills genbki.sh and Gen_fmgrtab.sh; we now rely entirely on
Perl scripts for those build steps. To avoid creating a Perl build dependency
where there was not one before, the output files generated by these scripts
are now treated as distprep targets, ie, they will be built and shipped in
tarballs. But you will need a reasonably modern Perl (probably at least
5.6) if you want to build from a CVS pull.
The changes to the MSVC build process are untested, and may well break ---
we'll soon find out from the buildfarm.
John Naylor, based on ideas from Robert Haas and others
probably got there via blind copy-and-paste from one of the legitimate
callers, so rearrange and comment that code a bit to make it clearer that
this isn't a necessary prerequisite to hash_create. Per observation
from Robert Haas.
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
relation rowtype OID into the relcache entries it builds. This ensures
that catcache copies of the relation tupdescs will be fully correct.
While the deficiency doesn't seem to have any effect in the current
sources, we have been bitten by not-quite-right catcache tupdescs before,
so it seems like a good idea to maintain the rule that they should be right.
possibility of shared-inval messages causing a relcache flush while it tries
to fill in missing data in preloaded relcache entries. There are actually
two distinct failure modes here:
1. The flush could delete the next-to-be-processed cache entry, causing
the subsequent hash_seq_search calls to go off into the weeds. This is
the problem reported by Michael Brown, and I believe it also accounts
for bug #5074. The simplest fix is to restart the hashtable scan after
we've read any new data from the catalogs. It appears that pre-8.4
branches have not suffered from this failure, because by chance there were
no other catalogs sharing the same hash chains with the catalogs that
RelationCacheInitializePhase2 had work to do for. However that's obviously
pretty fragile, and it seems possible that derivative versions with
additional system catalogs might be vulnerable, so I'm back-patching this
part of the fix anyway.
2. The flush could delete the *current* cache entry, in which case the
pointer to the newly-loaded data would end up being stored into an
already-deleted Relation struct. As long as it was still deleted, the only
consequence would be some leaked space in CacheMemoryContext. But it seems
possible that the Relation struct could already have been recycled, in
which case this represents a hard-to-reproduce clobber of cached data
structures, with unforeseeable consequences. The fix here is to pin the
entry while we work on it.
In passing, also change RelationCacheInitializePhase2 to Assert that
formrdesc() set up the relation's cached TupleDesc (rd_att) with the
correct type OID and hasoids values. This is more appropriate than
silently updating the values, because the original tupdesc might already
have been copied into the catcache. However this part of the patch is
not in HEAD because it fails due to some questionable recent changes in
formrdesc :-(. That will be cleaned up in a subsequent patch.
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.
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
temp relations; this is no more expensive than before, now that we have
pg_class.relistemp. Insert tests into bufmgr.c to prevent attempting
to fetch pages from nonlocal temp relations. This provides a low-level
defense against bugs-of-omission allowing temp pages to be loaded into shared
buffers, as in the contrib/pgstattuple problem reported by Stuart Bishop.
While at it, tweak a bunch of places to use new relcache tests (instead of
expensive probes into pg_namespace) to detect local or nonlocal temp tables.
relations (including a temp table's indexes and toast table/index), and
false for normal relations. For ease of checking, this commit just adds
the column and fills it correctly --- revising the relation access machinery
to use it will come separately.
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.
field needs to be included in equalRuleLocks() comparisons, else updates
will fail to propagate into relcache entries when they have positive
reference count (ie someone is using the relcache entry).
Per report from Alex Hunsaker.
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.
Tom Lane