Fujii Masao [Wed, 19 Nov 2014 10:10:04 +0000 (19:10 +0900)]
Fix bug in the test of file descriptor of current WAL file in pg_receivexlog.
In pg_receivexlog, in order to check whether the current WAL file is
being opened or not, its file descriptor has to be checked against -1
as an invalid value. But, oops, 7900e94 added the incorrect test
checking the descriptor against 1. This commit fixes that bug.
Fujii Masao [Wed, 19 Nov 2014 05:11:12 +0000 (14:11 +0900)]
Fix pg_receivexlog --slot so that it doesn't prevent the server shutdown.
When pg_receivexlog --slot is connecting to the server, at the shutdown
of the server, walsender keeps waiting for the last WAL record to be
replicated and flushed in pg_receivexlog. But previously pg_receivexlog
issued sync command only when WAL file was switched. So there was
the case where the last WAL was never flushed and walsender had to
keep waiting infinitely. This caused the server shutdown to get stuck.
pg_recvlogical handles this problem by calling fsync() when it receives
the request of immediate reply from the server. That is, at shutdown,
walsender sends the request, pg_recvlogical receives it, flushes the last
WAL record, and sends the flush location back to the server. Since
walsender can see that the last WAL record is successfully flushed, it can
exit cleanly.
This commit introduces the same logic as pg_recvlogical has,
to pg_receivexlog.
Back-patch to 9.4 where pg_receivexlog was changed so that it can use
the replication slot.
Original patch by Michael Paquier, rewritten by me.
Bug report by Furuya Osamu.
Tom Lane [Wed, 19 Nov 2014 02:36:39 +0000 (21:36 -0500)]
Don't require bleeding-edge timezone data in timestamptz regression test.
The regression test cases added in commits b2cbced9e et al depended in part
on the Russian timezone offset changes of Oct 2014. While this is of no
particular concern for a default Postgres build, it was possible for a
build using --with-system-tzdata to fail the tests if the system tzdata
database wasn't au courant. Bjorn Munch and Christoph Berg both complained
about this while packaging 9.4rc1, so we probably shouldn't insist on the
system tzdata being up-to-date. Instead, make an equivalent test using a
zone change that occurred in Venezuela in 2007. With this patch, the
regression tests should pass using any tzdata set from 2012 or later.
(I can't muster much sympathy for somebody using --with-system-tzdata
on a machine whose system tzdata is more than three years out-of-date.)
Tom Lane [Tue, 18 Nov 2014 20:51:45 +0000 (15:51 -0500)]
Update comments in find_typedef.
These comments don't seem to have been touched in a long time. Make them
describe the current implementation rather than what was here last century,
and be a bit more explicit about the unreferenced-typedefs issue.
Tom Lane [Tue, 18 Nov 2014 18:28:06 +0000 (13:28 -0500)]
Fix some bogus direct uses of realloc().
pg_dump/parallel.c was using realloc() directly with no error check.
While the odds of an actual failure here seem pretty low, Coverity
complains about it, so fix by using pg_realloc() instead.
While looking for other instances, I noticed a couple of places in
psql that hadn't gotten the memo about the availability of pg_realloc.
These aren't bugs, since they did have error checks, but verbosely
inconsistent code is not a good thing.
Back-patch as far as 9.3. 9.2 did not have pg_dump/parallel.c, nor
did it have pg_realloc available in all frontend code.
Simon Riggs [Tue, 18 Nov 2014 10:24:55 +0000 (10:24 +0000)]
Reduce btree scan overhead for < and > strategies
For <, <=, > and >= strategies, mark the first scan key
as already matched if scanning in an appropriate direction.
If index tuple contains no nulls we can skip the first
re-check for each tuple.
Author: Rajeev Rastogi
Reviewer: Haribabu Kommi
Rework of the code and comments by Simon Riggs
Simon Riggs [Mon, 17 Nov 2014 22:15:07 +0000 (22:15 +0000)]
Add pg_dump --snapshot option
Allows pg_dump to use a snapshot previously defined by a concurrent
session that has either used pg_export_snapshot() or obtained a
snapshot when creating a logical slot. When this option is used with
parallel pg_dump, the snapshot defined by this option is used and no
new snapshot is taken.
Fujii Masao [Mon, 17 Nov 2014 17:32:48 +0000 (02:32 +0900)]
Add --synchronous option to pg_receivexlog, for more reliable WAL writing.
Previously pg_receivexlog flushed WAL data only when WAL file was switched.
Then 3dad73e added -F option to pg_receivexlog so that users could control
how frequently sync commands were issued to WAL files. It also allowed users
to make pg_receivexlog flush WAL data immediately after writing by
specifying 0 in -F option. However feedback messages were not sent back
immediately even after a flush location was updated. So even if WAL data
was flushed in real time, the server could not see that for a while.
This commit removes -F option from and adds --synchronous to pg_receivexlog.
If --synchronous is specified, like the standby's wal receiver, pg_receivexlog
flushes WAL data as soon as there is WAL data which has not been flushed yet.
Then it sends back the feedback message identifying the latest flush location
to the server. This option is useful to make pg_receivexlog behave as sync
standby by using replication slot, for example.
Original patch by Furuya Osamu, heavily rewritten by me.
Reviewed by Heikki Linnakangas, Alvaro Herrera and Sawada Masahiko.
Tom Lane [Mon, 17 Nov 2014 17:08:02 +0000 (12:08 -0500)]
Update time zone data files to tzdata release 2014j.
DST law changes in the Turks & Caicos Islands (America/Grand_Turk) and
in Fiji. New zone Pacific/Bougainville for portions of Papua New Guinea.
Historical changes for Korea and Vietnam.
Fix WAL-logging of B-tree "unlink halfdead page" operation.
There was some confusion on how to record the case that the operation
unlinks the last non-leaf page in the branch being deleted.
_bt_unlink_halfdead_page set the "topdead" field in the WAL record to
the leaf page, but the redo routine assumed that it would be an invalid
block number in that case. This commit fixes _bt_unlink_halfdead_page to
do what the redo routine expected.
Alvaro Herrera [Mon, 17 Nov 2014 14:23:35 +0000 (11:23 -0300)]
Fix relpersistence setting in reindex_index
Buildfarm members with CLOBBER_CACHE_ALWAYS advised us that commit 85b506bbfc2937 was mistaken in setting the relpersistence value of the
index directly in the relcache entry, within reindex_index. The reason
for the failure is that an invalidation message that comes after mucking
with the relcache entry directly, but before writing it to the catalogs,
would cause the entry to become rebuilt in place from catalogs with the
old contents, losing the update.
Fix by passing the correct persistence value to
RelationSetNewRelfilenode instead; this routine also writes the updated
tuple to pg_class, avoiding the problem. Suggested by Tom Lane.
Simon Riggs [Sat, 15 Nov 2014 22:49:54 +0000 (22:49 +0000)]
Emit msg re skipping ANALYZE for absent inh tree
When checking a table that has an inheritance tree marked,
if no child tables remain, we skip ANALYZE. This patch emits
a message to show that the action has been skipped.
Alvaro Herrera [Sat, 15 Nov 2014 04:19:49 +0000 (01:19 -0300)]
Get rid of SET LOGGED indexes persistence kludge
This removes ATChangeIndexesPersistence() introduced by f41872d0c1239d36
which was too ugly to live for long. Instead, the correct persistence
marking is passed all the way down to reindex_index, so that the
transient relation built to contain the index relfilenode can
get marked correctly right from the start.
Author: Fabrízio de Royes Mello
Review and editorialization by Michael Paquier
and Álvaro Herrera
Andres Freund [Fri, 14 Nov 2014 17:22:12 +0000 (18:22 +0100)]
Fix initdb --sync-only to also sync tablespaces.
630cd14426dc added initdb --sync-only, for use by pg_upgrade, by just
exposing the existing fsync code. That's wrong, because initdb so far
had absolutely no reason to deal with tablespaces.
Fix --sync-only by additionally explicitly syncing each of the
tablespaces.
Backpatch to 9.3 where --sync-only was introduced.
Andres Freund [Fri, 14 Nov 2014 17:21:30 +0000 (18:21 +0100)]
Sync unlogged relations to disk after they have been reset.
Unlogged relations are only reset when performing a unclean
restart. That means they have to be synced to disk during clean
shutdowns. During normal processing that's achieved by registering a
buffer's file to be fsynced at the next checkpoint when flushed. But
ResetUnloggedRelations() doesn't go through the buffer manager, so
nothing will force reset relations to disk before the next shutdown
checkpoint.
So just make ResetUnloggedRelations() fsync the newly created main
forks to disk.
Andres Freund [Fri, 14 Nov 2014 17:20:59 +0000 (18:20 +0100)]
Ensure unlogged tables are reset even if crash recovery errors out.
Unlogged relations are reset at the end of crash recovery as they're
only synced to disk during a proper shutdown. Unfortunately that and
later steps can fail, e.g. due to running out of space. This reset
was, up to now performed after marking the database as having finished
crash recovery successfully. As out of space errors trigger a crash
restart that could lead to the situation that not all unlogged
relations are reset.
Once that happend usage of unlogged relations could yield errors like
"could not open file "...": No such file or directory". Luckily
clusters that show the problem can be fixed by performing a immediate
shutdown, and starting the database again.
To fix, just call ResetUnloggedRelations(UNLOGGED_RELATION_INIT)
earlier, before marking the database as having successfully recovered.
Tom Lane [Fri, 14 Nov 2014 22:19:29 +0000 (17:19 -0500)]
Document evaluation-order considerations for aggregate functions.
The SELECT reference page didn't really address the question of when
aggregate function evaluation occurs, nor did the "expression evaluation
rules" documentation mention that CASE can't be used to control whether
an aggregate gets evaluated or not. Improve that.
Per discussion of bug #11661. Original text by Marti Raudsepp and Michael
Paquier, rewritten significantly by me.
Stephen Frost [Fri, 14 Nov 2014 21:53:51 +0000 (16:53 -0500)]
Clean up includes from RLS patch
The initial patch for RLS mistakenly included headers associated with
the executor and planner bits in rewrite/rowsecurity.h. Per policy and
general good sense, executor headers should not be included in planner
headers or vice versa.
The include of execnodes.h was a mistaken holdover from previous
versions, while the include of relation.h was used for Relation's
definition, which should have been coming from utils/relcache.h. This
patch cleans these issues up, adds comments to the RowSecurityPolicy
struct and the RowSecurityConfigType enum, and changes Relation->rsdesc
to Relation->rd_rsdesc to follow Relation field naming convention.
Additionally, utils/rel.h was including rewrite/rowsecurity.h, which
wasn't a great idea since that was pulling in things not really needed
in utils/rel.h (which gets included in quite a few places). Instead,
use 'struct RowSecurityDesc' for the rd_rsdesc field and add comments
explaining why.
Lastly, add an include into access/nbtree/nbtsort.c for
utils/sortsupport.h, which was evidently missed due to the above mess.
Pointed out by Tom in 16970.1415838651@sss.pgh.pa.us; note that the
concerns regarding a similar situation in the custom-path commit still
need to be addressed.
Stephen Frost [Fri, 14 Nov 2014 20:16:01 +0000 (15:16 -0500)]
Revert change to ALTER TABLESPACE summary.
When ALTER TABLESPACE MOVE ALL was changed to be ALTER TABLE ALL IN
TABLESPACE, the ALTER TABLESPACE summary should have been adjusted back
to its original definition.
Alvaro Herrera [Fri, 14 Nov 2014 19:27:26 +0000 (16:27 -0300)]
Reduce disk footprint of brin regression test
Per complaint from Tom.
While at it, throw in some extra tests for nulls as well, and make sure
that the set of data we insert on the second round is not identical to
the first one. Both measures are intended to improve coverage of the
test.
Also uncomment the ON COMMIT DROP clause on the CREATE TEMP TABLE
commands. This doesn't have any effect for someone examining the
regression database after the tests are done, but it reduces clutter for
those that execute the script directly.
Alvaro Herrera [Fri, 14 Nov 2014 18:14:01 +0000 (15:14 -0300)]
Allow interrupting GetMultiXactIdMembers
This function has a loop which can lead to uninterruptible process
"stalls" (actually infinite loops) when some bugs are triggered. Avoid
that unpleasant situation by adding a check for interrupts in a place
that shouldn't degrade performance in the normal case.
Backpatch to 9.3. Older branches have an identical loop here, but the
aforementioned bugs are only a problem starting in 9.3 so there doesn't
seem to be any point in backpatching any further.
Andres Freund [Fri, 14 Nov 2014 16:04:44 +0000 (17:04 +0100)]
Move BufferGetBlockNumber() out of heap_page_is_all_visible()'s inner loop.
In some workloads BufferGetBlockNumber() shows up in profiles due to
the sheer number of calls to it (and because it causes cache
misses). The compiler can't move it out of the loop because it's a
full extern function call...
Andres Freund [Fri, 14 Nov 2014 15:58:00 +0000 (16:58 +0100)]
Add valgrind suppression for pg_atomic_init_u64.
pg_atomic_init_u64 (indirectly) uses compare/exchange to guarantee
atomic writes on platforms where compare/exchange is available, but
64bit writes aren't atomic (yes, those exist). That leads to a
harmless read of the initial value of variable.
Tom Lane [Thu, 13 Nov 2014 23:19:26 +0000 (18:19 -0500)]
Fix pg_dumpall to restore its ability to dump from ancient servers.
Fix breakage induced by commits d8d3d2a4f37f6df5d0118b7f5211978cca22091a
and 463f2625a5fb183b6a8925ccde98bb3889f921d9: pg_dumpall has crashed when
attempting to dump from pre-8.1 servers since then, due to faulty
construction of the query used for dumping roles from older servers.
The query was erroneous as of the earlier commit, but it wasn't exposed
unless you tried to use --binary-upgrade, which you presumably wouldn't
with a pre-8.1 server. However commit 463f2625a made it fail always.
In HEAD, also fix additional breakage induced in the same query by
commit 491c029dbc4206779cf659aa0ff986af7831d2ff, which evidently wasn't
tested against pre-8.1 servers either.
The bug is only latent in 9.1 because 463f2625a hadn't landed yet, but
it seems best to back-patch all branches containing the faulty query.
Andres Freund [Thu, 13 Nov 2014 18:06:43 +0000 (19:06 +0100)]
Fix and improve cache invalidation logic for logical decoding.
There are basically three situations in which logical decoding needs
to perform cache invalidation. During/After replaying a transaction
with catalog changes, when skipping a uninteresting transaction that
performed catalog changes and when erroring out while replaying a
transaction. Unfortunately these three cases were all done slightly
differently - partially because 8de3e410fa, which greatly simplifies
matters, got committed in the midst of the development of logical
decoding.
The actually problematic case was when logical decoding skipped
transaction commits (and thus processed invalidations). When used via
the SQL interface cache invalidation could access the catalog - bad,
because we didn't set up enough state to allow that correctly. It'd
not be hard to setup sufficient state, but the simpler solution is to
always perform cache invalidation outside a valid transaction.
Also make the different cache invalidation cases look as similar as
possible, to ease code review.
This fixes the assertion failure reported by Antonin Houska in 53EE02D9.7040702@gmail.com. The presented testcase has been expanded
into a regression test.
Backpatch to 9.4, where logical decoding was introduced.
Andres Freund [Thu, 13 Nov 2014 18:06:43 +0000 (19:06 +0100)]
Fix xmin/xmax horizon computation during logical decoding initialization.
When building the initial historic catalog snapshot there were
scenarios where snapbuild.c would use incorrect xmin/xmax values when
starting from a xl_running_xacts record. The values used were always a
bit suspect, but happened to be correct in the easy to test
cases. Notably the values used when the the initial snapshot was
computed while no other transactions were running were correct.
This is likely to be the cause of the occasional buildfarm failures on
animals markhor and tick; but it's quite possible to reproduce
problems without CLOBBER_CACHE_ALWAYS.
Backpatch to 9.4, where logical decoding was introduced.
Fix race condition between hot standby and restoring a full-page image.
There was a window in RestoreBackupBlock where a page would be zeroed out,
but not yet locked. If a backend pinned and locked the page in that window,
it saw the zeroed page instead of the old page or new page contents, which
could lead to missing rows in a result set, or errors.
To fix, replace RBM_ZERO with RBM_ZERO_AND_LOCK, which atomically pins,
zeroes, and locks the page, if it's not in the buffer cache already.
In stable branches, the old RBM_ZERO constant is renamed to RBM_DO_NOT_USE,
to avoid breaking any 3rd party extensions that might use RBM_ZERO. More
importantly, this avoids renumbering the other enum values, which would
cause even bigger confusion in extensions that use ReadBufferExtended, but
haven't been recompiled.
Backpatch to all supported versions; this has been racy since hot standby
was introduced.
Tom Lane [Wed, 12 Nov 2014 20:58:37 +0000 (15:58 -0500)]
Explicitly support the case that a plancache's raw_parse_tree is NULL.
This only happens if a client issues a Parse message with an empty query
string, which is a bit odd; but since it is explicitly called out as legal
by our FE/BE protocol spec, we'd probably better continue to allow it.
Fix by adding tests everywhere that the raw_parse_tree field is passed to
functions that don't or shouldn't accept NULL. Also make it clear in the
relevant comments that NULL is an expected case.
This reverts commits a73c9dbab0165b3395dfe8a44a7dfd16166963c4 and 2e9650cbcff8c8fb0d9ef807c73a44f241822eee, which fixed specific crash
symptoms by hacking things at what now seems to be the wrong end, ie the
callee functions. Making the callees allow NULL is superficially more
robust, but it's not always true that there is a defensible thing for the
callee to do in such cases. The caller has more context and is better
able to decide what the empty-query case ought to do.
Per followup discussion of bug #11335. Back-patch to 9.2. The code
before that is sufficiently different that it would require development
of a separate patch, which doesn't seem worthwhile for what is believed
to be an essentially cosmetic change.
Andres Freund [Wed, 12 Nov 2014 17:52:49 +0000 (18:52 +0100)]
Fix several weaknesses in slot and logical replication on-disk serialization.
Heikki noticed in 544E23C0.8090605@vmware.com that slot.c and
snapbuild.c were missing the FIN_CRC32 call when computing/checking
checksums of on disk files. That doesn't lower the the error detection
capabilities of the checksum, but is inconsistent with other usages.
In a followup mail Heikki also noticed that, contrary to a comment,
the 'version' and 'length' struct fields of replication slot's on disk
data where not covered by the checksum. That's not likely to lead to
actually missed corruption as those fields are cross checked with the
expected version and the actual file length. But it's wrong
nonetheless.
As fixing these issues makes existing on disk files unreadable, bump
the expected versions of on disk files for both slots and logical
decoding historic catalog snapshots. This means that loading old
files will fail with
ERROR: "replication slot file ... has unsupported version 1"
and
ERROR: "snapbuild state file ... has unsupported version 1 instead of
2" respectively. Given the low likelihood of anybody already using
these new features in a production setup that seems acceptable.
Fixing these issues made me notice that there's no regression test
covering the loading of historic snapshot from disk - so add one.
Backpatch to 9.4 where these features were introduced.
Andres Freund [Wed, 12 Nov 2014 17:52:49 +0000 (18:52 +0100)]
Add interrupt checks to contrib/pg_prewarm.
Currently the extension's pg_prewarm() function didn't check
interrupts once it started "warming" data. Since individual calls can
take a long while it's important for them to be interruptible.
Noah Misch [Wed, 12 Nov 2014 12:33:17 +0000 (07:33 -0500)]
Use just one database connection in the "tablespace" test.
On Windows, DROP TABLESPACE has a race condition when run concurrently
with other processes having opened files in the tablespace. This led to
a rare failure on buildfarm member frogmouth. Back-patch to 9.4, where
the reconnection was introduced.
Tom Lane [Tue, 11 Nov 2014 22:00:11 +0000 (17:00 -0500)]
Fix dependency searching for case where column is visited before table.
When the recursive search in dependency.c visits a column and then later
visits the whole table containing the column, it needs to propagate the
drop-context flags for the table to the existing target-object entry for
the column. Otherwise we might refuse the DROP (if not CASCADE) on the
incorrect grounds that there was no automatic drop pathway to the column.
Remarkably, this has not been reported before, though it's possible at
least when an extension creates both a datatype and a table using that
datatype.
Rather than just marking the column as allowed to be dropped, it might
seem good to skip the DROP COLUMN step altogether, since the later DROP
of the table will surely get the job done. The problem with that is that
the datatype would then be dropped before the table (since the whole
situation occurred because we visited the datatype, and then recursed to
the dependent column, before visiting the table). That seems pretty risky,
and the case is rare enough that it doesn't seem worth expending a lot of
effort or risk to make the drops happen in a safe order. So we just play
dumb and delete the column separately according to the existing drop
ordering rules.
Per report from Petr Jelinek, though this is different from his proposed
patch.
Back-patch to 9.1, where extensions were introduced. There's currently
no evidence that such cases can arise before 9.1, and in any case we would
also need to back-patch cb5c2ba2d82688d29b5902d86b993a54355cad4d to 9.0
if we wanted to back-patch this.
Fujii Masao [Tue, 11 Nov 2014 12:08:21 +0000 (21:08 +0900)]
Add GUC and storage parameter to set the maximum size of GIN pending list.
Previously the maximum size of GIN pending list was controlled only by
work_mem. But the reasonable value of work_mem and the reasonable size
of the list are basically not the same, so it was not appropriate to
control both of them by only one GUC, i.e., work_mem. This commit
separates new GUC, pending_list_cleanup_size, from work_mem to allow
users to control only the size of the list.
Also this commit adds pending_list_cleanup_size as new storage parameter
to allow users to specify the size of the list per index. This is useful,
for example, when users want to increase the size of the list only for
the GIN index which can be updated heavily, and decrease it otherwise.
I missed an additional colon in previous patch. Oops. to make that mistake
less likely in the future, add comments as placeholders for unused inputs
and outputs in inline assembly.
Alvaro Herrera [Mon, 10 Nov 2014 21:13:49 +0000 (18:13 -0300)]
BRIN: fix bug in xlog backup block counting
The code that generates the BRIN_XLOG_UPDATE removes the buffer
reference when the page that's target for the updated tuple is freshly
initialized. This is a pretty usual optimization, but was breaking the
case where the revmap buffer, which is referenced in the same WAL
record, is getting a backup block: the replay code was using backup
block index 1, which is not valid when the update target buffer gets
pruned; the revmap buffer gets assigned 0 instead. Make sure to use the
correct backup block index for revmap when replaying.
Robert Haas [Mon, 10 Nov 2014 20:19:56 +0000 (15:19 -0500)]
Fix potential NULL-pointer dereference.
Commit 2781b4bea7db357be59f9a5fd73ca1eb12ff5a79 arranged to defer
the setup of after-trigger-related data structures, but
AfterTriggerPendingOnRel didn't get the memo.
Tom Lane [Mon, 10 Nov 2014 20:21:09 +0000 (15:21 -0500)]
Ensure that RowExprs and whole-row Vars produce the expected column names.
At one time it wasn't terribly important what column names were associated
with the fields of a composite Datum, but since the introduction of
operations like row_to_json(), it's important that looking up the rowtype
ID embedded in the Datum returns the column names that users would expect.
That did not work terribly well before this patch: you could get the column
names of the underlying table, or column aliases from any level of the
query, depending on minor details of the plan tree. You could even get
totally empty field names, which is disastrous for cases like row_to_json().
To fix this for whole-row Vars, look to the RTE referenced by the Var, and
make sure its column aliases are applied to the rowtype associated with
the result Datums. This is a tad scary because we might have to return
a transient RECORD type even though the Var is declared as having some
named rowtype. In principle it should be all right because the record
type will still be physically compatible with the named rowtype; but
I had to weaken one Assert in ExecEvalConvertRowtype, and there might be
third-party code containing similar assumptions.
Similarly, RowExprs have to be willing to override the column names coming
from a named composite result type and produce a RECORD when the column
aliases visible at the site of the RowExpr differ from the underlying
table's column names.
In passing, revert the decision made in commit 398f70ec070fe601 to add
an alias-list argument to ExecTypeFromExprList: better to provide that
functionality in a separate function. This also reverts most of the code
changes in d68581483564ec0f, which we don't need because we're no longer
depending on the tupdesc found in the child plan node's result slot to be
blessed.
Back-patch to 9.4, but not earlier, since this solution changes the results
in some cases that users might not have realized were buggy. We'll apply a
more restricted form of this patch in older branches.
Alvaro Herrera [Mon, 10 Nov 2014 18:56:08 +0000 (15:56 -0300)]
Further code and wording tweaks in BRIN
Besides a couple of typo fixes, per David Rowley, Thom Brown, and Amit
Langote, and mentions of BRIN in the general CREATE INDEX page again per
David, this includes silencing MSVC compiler warnings (thanks Microsoft)
and an additional variable initialization per Coverity scanner.
Alvaro Herrera [Sat, 8 Nov 2014 03:31:03 +0000 (00:31 -0300)]
Fix some coding issues in BRIN
Reported by David Rowley: variadic macros are a problem. Get rid of
them using a trick suggested by Tom Lane: add extra parentheses where
needed. In the future we might decide we don't need the calls at all
and remove them, but it seems appropriate to keep them while this code
is still new.
Also from David Rowley: brininsert() was trying to use a variable before
initializing it. Fix by moving the brin_form_tuple call (which
initializes the variable) to within the locked section.
Reported by Peter Eisentraut: can't use "new" as a struct member name,
because C++ compilers will choke on it, as reported by cpluspluscheck.
pg_basebackup: Adjust tests for long file name issues
Work around accidental test failures because the working directory path
is too long by creating a temporary directory in the (hopefully shorter)
system location, symlinking that to the working directory, and creating
the tablespaces using the shorter path.
Robert Haas [Fri, 7 Nov 2014 22:26:02 +0000 (17:26 -0500)]
Introduce custom path and scan providers.
This allows extension modules to define their own methods for
scanning a relation, and get the core code to use them. It's
unclear as yet how much use this capability will find, but we
won't find out if we never commit it.
KaiGai Kohei, reviewed at various times and in various levels
of detail by Shigeru Hanada, Tom Lane, Andres Freund, Álvaro
Herrera, and myself.
Now that the backup blocks are appended to the WAL record in xloginsert.c,
XLogInsert doesn't see them anymore and cannot remove them from the version
reconstructed for xlog_outdesc. This makes running with wal_debug=on more
expensive, as we now make (unnecessary) temporary copies of the backup
blocks, but it doesn't seem worth convoluting the code to keep that
optimization.
Alvaro Herrera [Fri, 7 Nov 2014 20:05:26 +0000 (17:05 -0300)]
Fix serial schedule
Test misc depends on brin, but it was earlier in the serial schedule
file. I didn't notice this because I only run the parallel schedule,
but the buildfarm exposed my folly ...
Alvaro Herrera [Fri, 7 Nov 2014 19:38:14 +0000 (16:38 -0300)]
BRIN: Block Range Indexes
BRIN is a new index access method intended to accelerate scans of very
large tables, without the maintenance overhead of btrees or other
traditional indexes. They work by maintaining "summary" data about
block ranges. Bitmap index scans work by reading each summary tuple and
comparing them with the query quals; all pages in the range are returned
in a lossy TID bitmap if the quals are consistent with the values in the
summary tuple, otherwise not. Normal index scans are not supported
because these indexes do not store TIDs.
As new tuples are added into the index, the summary information is
updated (if the block range in which the tuple is added is already
summarized) or not; in the latter case, a subsequent pass of VACUUM or
the brin_summarize_new_values() function will create the summary
information.
For data types with natural 1-D sort orders, the summary info consists
of the maximum and the minimum values of each indexed column within each
page range. This type of operator class we call "Minmax", and we
supply a bunch of them for most data types with B-tree opclasses.
Since the BRIN code is generalized, other approaches are possible for
things such as arrays, geometric types, ranges, etc; even for things
such as enum types we could do something different than minmax with
better results. In this commit I only include minmax.
Catalog version bumped due to new builtin catalog entries.
There's more that could be done here, but this is a good step forwards.
Loosely based on ideas from Simon Riggs; code mostly by Álvaro Herrera,
with contribution by Heikki Linnakangas.
Patch reviewed by: Amit Kapila, Heikki Linnakangas, Robert Haas.
Testing help from Jeff Janes, Erik Rijkers, Emanuel Calvo.
PS:
The research leading to these results has received funding from the
European Union's Seventh Framework Programme (FP7/2007-2013) under
grant agreement n° 318633.
The code that generated a record to clear the F_TUPLES_DELETED flag hasn't
existed since we got rid of old-style VACUUM FULL. I kept the code that sets
the flag, although it's not used for anything anymore, because it might
still be interesting information for debugging purposes that some tuples
have been deleted from a page.
Likewise, the code to turn the root page from non-leaf to leaf page was
removed when we got rid of old-style VACUUM FULL. Remove the code to replay
that action, too.
Tom Lane [Fri, 7 Nov 2014 01:52:40 +0000 (20:52 -0500)]
Cope with more than 64K phrases in a thesaurus dictionary.
dict_thesaurus stored phrase IDs in uint16 fields, so it would get confused
and even crash if there were more than 64K entries in the configuration
file. It turns out to be basically free to widen the phrase IDs to uint32,
so let's just do so.
This was complained of some time ago by David Boutin (in bug #7793);
he later submitted an informal patch but it was never acted on.
We now have another complaint (bug #11901 from Luc Ouellette) so it's
time to make something happen.
This is basically Boutin's patch, but for future-proofing I also added a
defense against too many words per phrase. Note that we don't need any
explicit defense against overflow of the uint32 counters, since before that
happens we'd hit array allocation sizes that repalloc rejects.
Back-patch to all supported branches because of the crash risk.
Tom Lane [Thu, 6 Nov 2014 16:41:06 +0000 (11:41 -0500)]
Fix normalization of numeric values in JSONB GIN indexes.
The default JSONB GIN opclass (jsonb_ops) converts numeric data values
to strings for storage in the index. It must ensure that numeric values
that would compare equal (such as 12 and 12.00) produce identical strings,
else index searches would have behavior different from regular JSONB
comparisons. Unfortunately the function charged with doing this was
completely wrong: it could reduce distinct numeric values to the same
string, or reduce equivalent numeric values to different strings. The
former type of error would only lead to search inefficiency, but the
latter type of error would cause index entries that should be found by
a search to not be found.
Repairing this bug therefore means that it will be necessary for 9.4 beta
testers to reindex GIN jsonb_ops indexes, if they care about getting
correct results from index searches involving numeric data values within
the comparison JSONB object.
Fujii Masao [Thu, 6 Nov 2014 12:24:40 +0000 (21:24 +0900)]
Prevent the unnecessary creation of .ready file for the timeline history file.
Previously .ready file was created for the timeline history file at the end
of an archive recovery even when WAL archiving was not enabled.
This creation is unnecessary and causes .ready file to remain infinitely.
This commit changes an archive recovery so that it creates .ready file for
the timeline history file only when WAL archiving is enabled.
Move the backup-block logic from XLogInsert to a new file, xloginsert.c.
xlog.c is huge, this makes it a little bit smaller, which is nice. Functions
related to putting together the WAL record are in xloginsert.c, and the
lower level stuff for managing WAL buffers and such are in xlog.c.
Also move the definition of XLogRecord to a separate header file. This
causes churn in the #includes of all the files that write WAL records, and
redo routines, but it avoids pulling in xlog.h into most places.
Reviewed by Michael Paquier, Alvaro Herrera, Andres Freund and Amit Kapila.
Tom Lane [Thu, 6 Nov 2014 00:35:23 +0000 (19:35 -0500)]
Remove the last vestige of server-side autocommit.
Long ago we briefly had an "autocommit" GUC that turned server-side
autocommit on and off. That behavior was removed in 7.4 after concluding
that it broke far too much client-side logic, and making clients cope with
both behaviors was impractical. But the GUC variable was left behind, so
as not to break any client code that might be trying to read its value.
Enough time has now passed that we should remove the GUC completely.
Whatever vestigial backwards-compatibility benefit it had is outweighed by
the risk of confusion for newbies who assume it ought to do something,
as per a recent complaint from Wolfgang Wilhelm.
In passing, adjust what seemed to me a rather confusing documentation
reference to libpq's autocommit behavior. libpq as such knows nothing
about autocommit, so psql is probably what was meant.
Obviously, every translation unit should not be declaring this
separately. It needs to be PGDLLIMPORT as well, to avoid breaking
third-party code that uses any of the functions that the commit
mentioned above changed to macros.
Tom Lane [Wed, 5 Nov 2014 16:44:06 +0000 (11:44 -0500)]
Make CREATE TYPE print warnings if a datatype's I/O functions are volatile.
This is a followup to commit 43ac12c6e6e397fd9142ed908447eba32d3785b2,
which added regression tests checking that I/O functions of built-in
types are not marked volatile. Complaining in CREATE TYPE should push
developers of add-on types to fix any misdeclared functions in their
types. It's just a warning not an error, to avoid creating upgrade
problems for what might be just cosmetic mis-markings.
Aside from adding the warning code, fix a number of types that were
sloppily created in the regression tests.
Tom Lane [Wed, 5 Nov 2014 16:34:11 +0000 (11:34 -0500)]
Fix volatility markings of some contrib I/O functions.
In general, datatype I/O functions are supposed to be immutable or at
worst stable. Some contrib I/O functions were, through oversight, not
marked with any volatility property at all, which made them VOLATILE.
Since (most of) these functions actually behave immutably, the erroneous
marking isn't terribly harmful; but it can be user-visible in certain
circumstances, as per a recent bug report from Joe Van Dyk in which a
cast to text was disallowed in an expression index definition.
To fix, just adjust the declarations in the extension SQL scripts. If we
were being very fussy about this, we'd bump the extension version numbers,
but that seems like more trouble (for both developers and users) than the
problem is worth.
A fly in the ointment is that chkpass_in actually is volatile, because
of its use of random() to generate a fresh salt when presented with a
not-yet-encrypted password. This is bad because of the general assumption
that I/O functions aren't volatile: the consequence is that records or
arrays containing chkpass elements may have input behavior a bit different
from a bare chkpass column. But there seems no way to fix this without
breaking existing usage patterns for chkpass, and the consequences of the
inconsistency don't seem bad enough to justify that. So for the moment,
just document it in a comment.
Since we're not bumping version numbers, there seems no harm in
back-patching these fixes; at least future installations will get the
functions marked correctly.
Tom Lane [Tue, 4 Nov 2014 18:24:06 +0000 (13:24 -0500)]
Drop no-longer-needed buffers during ALTER DATABASE SET TABLESPACE.
The previous coding assumed that we could just let buffers for the
database's old tablespace age out of the buffer arena naturally.
The folly of that is exposed by bug #11867 from Marc Munro: the user could
later move the database back to its original tablespace, after which any
still-surviving buffers would match lookups again and appear to contain
valid data. But they'd be missing any changes applied while the database
was in the new tablespace.
This has been broken since ALTER SET TABLESPACE was introduced, so
back-patch to all supported branches.
The old algorithm was found to not be the usual CRC-32 algorithm, used by
Ethernet et al. We were using a non-reflected lookup table with code meant
for a reflected lookup table. That's a strange combination that AFAICS does
not correspond to any bit-wise CRC calculation, which makes it difficult to
reason about its properties. Although it has worked well in practice, seems
safer to use a well-known algorithm.
Since we're changing the algorithm anyway, we might as well choose a
different polynomial. The Castagnoli polynomial has better error-correcting
properties than the traditional CRC-32 polynomial, even if we had
implemented it correctly. Another reason for picking that is that some new
CPUs have hardware support for calculating CRC-32C, but not CRC-32, let
alone our strange variant of it. This patch doesn't add any support for such
hardware, but a future patch could now do that.
The old algorithm is kept around for tsquery and pg_trgm, which use the
values in indexes that need to remain compatible so that pg_upgrade works.
While we're at it, share the old lookup table for CRC-32 calculation
between hstore, ltree and core. They all use the same table, so might as
well.
Noah Misch [Mon, 3 Nov 2014 02:43:25 +0000 (21:43 -0500)]
Make ECPG test programs depend on "ecpg$(X)", not "ecpg".
Cygwin builds require this of dependencies pertaining to pattern rules.
On Cygwin, stat("foo") in the absence of a file with that exact name can
locate foo.exe. While GNU make uses stat() for dependencies of ordinary
rules, it uses readdir() to assess dependencies of pattern rules.
Therefore, a pattern rule dependency should match any underlying file
name exactly. Back-patch to 9.4, where the dependency was introduced.
Robert Haas [Fri, 31 Oct 2014 16:02:40 +0000 (12:02 -0400)]
Support frontend-backend protocol communication using a shm_mq.
A background worker can use pq_redirect_to_shm_mq() to direct protocol
that would normally be sent to the frontend to a shm_mq so that another
process may read them.
The receiving process may use pq_parse_errornotice() to parse an
ErrorResponse or NoticeResponse from the background worker and, if
it wishes, ThrowErrorData() to propagate the error (with or without
further modification).