Merge pull request #155 from postgres/master

Sync Fork from Upstream Repo

Author: sthagen

doc/src/sgml/backup.sgml

@@ -1520,11 +1520,11 @@ tar -rf /var/lib/pgsql/backup.tar /var/lib/pgsql/archive/
       If archive storage size is a concern, you can use
       <application>gzip</application> to compress the archive files:
 <programlisting>
-archive_command = 'gzip &lt; %p &gt; /var/lib/pgsql/archive/%f'
+archive_command = 'gzip &lt; %p &gt; /mnt/server/archivedir/%f.gz'
 </programlisting>
       You will then need to use <application>gunzip</application> during recovery:
 <programlisting>
-restore_command = 'gunzip &lt; /mnt/server/archivedir/%f &gt; %p'
+restore_command = 'gunzip &lt; /mnt/server/archivedir/%f.gz &gt; %p'
 </programlisting>
      </para>
     </sect3>

doc/src/sgml/func.sgml

@@ -26238,14 +26238,14 @@ postgres=# SELECT * FROM pg_walfile_name_offset(pg_stop_backup());
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
-         <primary>pg_collation_current_version</primary>
+         <primary>pg_collation_actual_version</primary>
         </indexterm>
-        <function>pg_collation_current_version</function> ( <type>oid</type> )
+        <function>pg_collation_actual_version</function> ( <type>oid</type> )
         <returnvalue>text</returnvalue>
        </para>
        <para>
-        Returns the version of the collation object as reported by the ICU
-        library or operating system.  <literal>null</literal> is returned
+        Returns the actual version of the collation object as it is currently
+        installed in the operating system.  <literal>null</literal> is returned
         on operating systems where <productname>PostgreSQL</productname>
         doesn't have support for versions.
        </para></entry>

src/backend/access/gin/ginvacuum.c

@@ -231,6 +231,7 @@ ginDeletePage(GinVacuumState *gvs, BlockNumber deleteBlkno, BlockNumber leftBlkn
 
 	END_CRIT_SECTION();
 
+	gvs->result->pages_newly_deleted++;
 	gvs->result->pages_deleted++;
 }
 

src/backend/access/gist/gistvacuum.c

@@ -133,9 +133,21 @@ gistvacuumscan(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
 	MemoryContext oldctx;
 
 	/*
-	 * Reset counts that will be incremented during the scan; needed in case
-	 * of multiple scans during a single VACUUM command.
+	 * Reset fields that track information about the entire index now.  This
+	 * avoids double-counting in the case where a single VACUUM command
+	 * requires multiple scans of the index.
+	 *
+	 * Avoid resetting the tuples_removed and pages_newly_deleted fields here,
+	 * since they track information about the VACUUM command, and so must last
+	 * across each call to gistvacuumscan().
+	 *
+	 * (Note that pages_free is treated as state about the whole index, not
+	 * the current VACUUM.  This is appropriate because RecordFreeIndexPage()
+	 * calls are idempotent, and get repeated for the same deleted pages in
+	 * some scenarios.  The point for us is to track the number of recyclable
+	 * pages in the index at the end of the VACUUM command.)
 	 */
+	stats->num_pages = 0;
 	stats->estimated_count = false;
 	stats->num_index_tuples = 0;
 	stats->pages_deleted = 0;
@@ -281,8 +293,8 @@ gistvacuumpage(GistVacState *vstate, BlockNumber blkno, BlockNumber orig_blkno)
 	{
 		/* Okay to recycle this page */
 		RecordFreeIndexPage(rel, blkno);
-		vstate->stats->pages_free++;
 		vstate->stats->pages_deleted++;
+		vstate->stats->pages_free++;
 	}
 	else if (GistPageIsDeleted(page))
 	{
@@ -636,6 +648,7 @@ gistdeletepage(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
 	/* mark the page as deleted */
 	MarkBufferDirty(leafBuffer);
 	GistPageSetDeleted(leafPage, txid);
+	stats->pages_newly_deleted++;
 	stats->pages_deleted++;
 
 	/* remove the downlink from the parent */

src/backend/access/heap/vacuumlazy.c

@@ -2521,9 +2521,11 @@ lazy_cleanup_index(Relation indrel,
 						(*stats)->num_index_tuples,
 						(*stats)->num_pages),
 				 errdetail("%.0f index row versions were removed.\n"
-						   "%u index pages have been deleted, %u are currently reusable.\n"
+						   "%u index pages were newly deleted.\n"
+						   "%u index pages are currently deleted, of which %u are currently reusable.\n"
 						   "%s.",
 						   (*stats)->tuples_removed,
+						   (*stats)->pages_newly_deleted,
 						   (*stats)->pages_deleted, (*stats)->pages_free,
 						   pg_rusage_show(&ru0))));
 	}

src/backend/access/nbtree/nbtpage.c

@@ -50,7 +50,7 @@ static bool _bt_mark_page_halfdead(Relation rel, Buffer leafbuf,
 static bool _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf,
 									 BlockNumber scanblkno,
 									 bool *rightsib_empty,
-									 uint32 *ndeleted);
+									 BTVacState *vstate);
 static bool _bt_lock_subtree_parent(Relation rel, BlockNumber child,
 									BTStack stack,
 									Buffer *subtreeparent,
@@ -1760,28 +1760,31 @@ _bt_rightsib_halfdeadflag(Relation rel, BlockNumber leafrightsib)
  * should never pass a buffer containing an existing deleted page here.  The
  * lock and pin on caller's buffer will be dropped before we return.
  *
- * Returns the number of pages successfully deleted (zero if page cannot
- * be deleted now; could be more than one if parent or right sibling pages
- * were deleted too).  Note that this does not include pages that we delete
- * that the btvacuumscan scan has yet to reach; they'll get counted later
- * instead.
+ * Maintains bulk delete stats for caller, which are taken from vstate.  We
+ * need to cooperate closely with caller here so that whole VACUUM operation
+ * reliably avoids any double counting of subsidiary-to-leafbuf pages that we
+ * delete in passing.  If such pages happen to be from a block number that is
+ * ahead of the current scanblkno position, then caller is expected to count
+ * them directly later on.  It's simpler for us to understand caller's
+ * requirements than it would be for caller to understand when or how a
+ * deleted page became deleted after the fact.
  *
  * NOTE: this leaks memory.  Rather than trying to clean up everything
  * carefully, it's better to run it in a temp context that can be reset
  * frequently.
  */
-uint32
-_bt_pagedel(Relation rel, Buffer leafbuf)
+void
+_bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
 {
-	uint32		ndeleted = 0;
 	BlockNumber rightsib;
 	bool		rightsib_empty;
 	Page		page;
 	BTPageOpaque opaque;
 
 	/*
 	 * Save original leafbuf block number from caller.  Only deleted blocks
-	 * that are <= scanblkno get counted in ndeleted return value.
+	 * that are <= scanblkno are added to bulk delete stat's pages_deleted
+	 * count.
 	 */
 	BlockNumber scanblkno = BufferGetBlockNumber(leafbuf);
 
@@ -1843,7 +1846,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 										 RelationGetRelationName(rel))));
 
 			_bt_relbuf(rel, leafbuf);
-			return ndeleted;
+			return;
 		}
 
 		/*
@@ -1873,7 +1876,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 			Assert(!P_ISHALFDEAD(opaque));
 
 			_bt_relbuf(rel, leafbuf);
-			return ndeleted;
+			return;
 		}
 
 		/*
@@ -1922,8 +1925,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 				if (_bt_leftsib_splitflag(rel, leftsib, leafblkno))
 				{
 					ReleaseBuffer(leafbuf);
-					Assert(ndeleted == 0);
-					return ndeleted;
+					return;
 				}
 
 				/* we need an insertion scan key for the search, so build one */
@@ -1964,7 +1966,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 			if (!_bt_mark_page_halfdead(rel, leafbuf, stack))
 			{
 				_bt_relbuf(rel, leafbuf);
-				return ndeleted;
+				return;
 			}
 		}
 
@@ -1979,7 +1981,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 		{
 			/* Check for interrupts in _bt_unlink_halfdead_page */
 			if (!_bt_unlink_halfdead_page(rel, leafbuf, scanblkno,
-										  &rightsib_empty, &ndeleted))
+										  &rightsib_empty, vstate))
 			{
 				/*
 				 * _bt_unlink_halfdead_page should never fail, since we
@@ -1990,7 +1992,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 				 * lock and pin on leafbuf for us.
 				 */
 				Assert(false);
-				return ndeleted;
+				return;
 			}
 		}
 
@@ -2026,8 +2028,6 @@ _bt_pagedel(Relation rel, Buffer leafbuf)
 
 		leafbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
 	}
-
-	return ndeleted;
 }
 
 /*
@@ -2262,9 +2262,10 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
  */
 static bool
 _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
-						 bool *rightsib_empty, uint32 *ndeleted)
+						 bool *rightsib_empty, BTVacState *vstate)
 {
 	BlockNumber leafblkno = BufferGetBlockNumber(leafbuf);
+	IndexBulkDeleteResult *stats = vstate->stats;
 	BlockNumber leafleftsib;
 	BlockNumber leafrightsib;
 	BlockNumber target;
@@ -2674,12 +2675,17 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
 		_bt_relbuf(rel, buf);
 
 	/*
-	 * If btvacuumscan won't revisit this page in a future btvacuumpage call
-	 * and count it as deleted then, we count it as deleted by current
-	 * btvacuumpage call
+	 * Maintain pages_newly_deleted, which is simply the number of pages
+	 * deleted by the ongoing VACUUM operation.
+	 *
+	 * Maintain pages_deleted in a way that takes into account how
+	 * btvacuumpage() will count deleted pages that have yet to become
+	 * scanblkno -- only count page when it's not going to get that treatment
+	 * later on.
 	 */
+	stats->pages_newly_deleted++;
 	if (target <= scanblkno)
-		(*ndeleted)++;
+		stats->pages_deleted++;
 
 	return true;
 }

src/backend/access/nbtree/nbtree.c

@@ -38,17 +38,6 @@
 #include "utils/memutils.h"
 
 
-/* Working state needed by btvacuumpage */
-typedef struct
-{
-	IndexVacuumInfo *info;
-	IndexBulkDeleteResult *stats;
-	IndexBulkDeleteCallback callback;
-	void	   *callback_state;
-	BTCycleId	cycleid;
-	MemoryContext pagedelcontext;
-} BTVacState;
-
 /*
  * BTPARALLEL_NOT_INITIALIZED indicates that the scan has not started.
  *
@@ -1016,9 +1005,9 @@ btvacuumscan(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
 	 * avoids double-counting in the case where a single VACUUM command
 	 * requires multiple scans of the index.
 	 *
-	 * Avoid resetting the tuples_removed field here, since it tracks
-	 * information about the VACUUM command, and so must last across each call
-	 * to btvacuumscan().
+	 * Avoid resetting the tuples_removed and pages_newly_deleted fields here,
+	 * since they track information about the VACUUM command, and so must last
+	 * across each call to btvacuumscan().
 	 *
 	 * (Note that pages_free is treated as state about the whole index, not
 	 * the current VACUUM.  This is appropriate because RecordFreeIndexPage()
@@ -1237,11 +1226,13 @@ btvacuumpage(BTVacState *vstate, BlockNumber scanblkno)
 	}
 	else if (P_ISHALFDEAD(opaque))
 	{
+		/* Half-dead leaf page (from interrupted VACUUM) -- finish deleting */
+		attempt_pagedel = true;
+
 		/*
-		 * Half-dead leaf page.  Try to delete now.  Might update
-		 * pages_deleted below.
+		 * _bt_pagedel() will increment both pages_newly_deleted and
+		 * pages_deleted stats in all cases (barring corruption)
 		 */
-		attempt_pagedel = true;
 	}
 	else if (P_ISLEAF(opaque))
 	{
@@ -1451,12 +1442,12 @@ btvacuumpage(BTVacState *vstate, BlockNumber scanblkno)
 		oldcontext = MemoryContextSwitchTo(vstate->pagedelcontext);
 
 		/*
-		 * We trust the _bt_pagedel return value because it does not include
-		 * any page that a future call here from btvacuumscan is expected to
-		 * count.  There will be no double-counting.
+		 * _bt_pagedel maintains the bulk delete stats on our behalf;
+		 * pages_newly_deleted and pages_deleted are likely to be incremented
+		 * during call
 		 */
 		Assert(blkno == scanblkno);
-		stats->pages_deleted += _bt_pagedel(rel, buf);
+		_bt_pagedel(rel, buf, vstate);
 
 		MemoryContextSwitchTo(oldcontext);
 		/* pagedel released buffer, so we shouldn't */

src/backend/access/spgist/spgvacuum.c

@@ -891,6 +891,7 @@ spgvacuumscan(spgBulkDeleteState *bds)
 
 	/* Report final stats */
 	bds->stats->num_pages = num_pages;
+	bds->stats->pages_newly_deleted = bds->stats->pages_deleted;
 	bds->stats->pages_free = bds->stats->pages_deleted;
 }
 

src/backend/commands/collationcmds.c

@@ -268,7 +268,7 @@ IsThereCollationInNamespace(const char *collname, Oid nspOid)
 }
 
 Datum
-pg_collation_current_version(PG_FUNCTION_ARGS)
+pg_collation_actual_version(PG_FUNCTION_ARGS)
 {
 	Oid			collid = PG_GETARG_OID(0);
 	char	   *version;

src/backend/parser/parse_cte.c

@@ -730,15 +730,15 @@ makeDependencyGraphWalker(Node *node, CteState *cstate)
 				 * In the non-RECURSIVE case, query names are visible to the
 				 * WITH items after them and to the main query.
 				 */
-				ListCell   *cell1;
-
 				cstate->innerwiths = lcons(NIL, cstate->innerwiths);
-				cell1 = list_head(cstate->innerwiths);
 				foreach(lc, stmt->withClause->ctes)
 				{
 					CommonTableExpr *cte = (CommonTableExpr *) lfirst(lc);
+					ListCell   *cell1;
 
 					(void) makeDependencyGraphWalker(cte->ctequery, cstate);
+					/* note that recursion could mutate innerwiths list */
+					cell1 = list_head(cstate->innerwiths);
 					lfirst(cell1) = lappend((List *) lfirst(cell1), cte);
 				}
 				(void) raw_expression_tree_walker(node,
@@ -1006,15 +1006,15 @@ checkWellFormedRecursionWalker(Node *node, CteState *cstate)
 				 * In the non-RECURSIVE case, query names are visible to the
 				 * WITH items after them and to the main query.
 				 */
-				ListCell   *cell1;
-
 				cstate->innerwiths = lcons(NIL, cstate->innerwiths);
-				cell1 = list_head(cstate->innerwiths);
 				foreach(lc, stmt->withClause->ctes)
 				{
 					CommonTableExpr *cte = (CommonTableExpr *) lfirst(lc);
+					ListCell   *cell1;
 
 					(void) checkWellFormedRecursionWalker(cte->ctequery, cstate);
+					/* note that recursion could mutate innerwiths list */
+					cell1 = list_head(cstate->innerwiths);
 					lfirst(cell1) = lappend((List *) lfirst(cell1), cte);
 				}
 				checkWellFormedSelectStmt(stmt, cstate);