Don't clear btpo_cycleid during _bt_vacuum_one_page.

tglsfdc · tglsfdc · commit 7b6a37f5ec2e · 2012-02-21T15:03:56.000-05:00
When "vacuuming" a single btree page by removing LP_DEAD tuples, we are not actually within a vacuum operation, but rather in an ordinary insertion process that could well be running concurrently with a vacuum. So clearing the cycleid is incorrect, and could cause the concurrent vacuum to miss removing tuples that it needs to remove. This is a longstanding bug introduced by commit e628464 of 2006-07-25. I believe it explains Maxim Boguk's recent report of index corruption, and probably some other previously unexplained reports. In 9.0 and up this is a one-line fix; before that we need to introduce a flag to tell _bt_delitems what to do.
diff --git a/src/backend/access/nbtree/nbtinsert.c b/src/backend/access/nbtree/nbtinsert.c
@@ -1978,7 +1978,7 @@ _bt_vacuum_one_page(Relation rel, Buffer buffer)
 	}
 
 	if (ndeletable > 0)
-		_bt_delitems(rel, buffer, deletable, ndeletable);
+		_bt_delitems(rel, buffer, deletable, ndeletable, false);
 
 	/*
 	 * Note: if we didn't find any LP_DEAD items, then the page's
diff --git a/src/backend/access/nbtree/nbtpage.c b/src/backend/access/nbtree/nbtpage.c
@@ -656,7 +656,8 @@ _bt_page_recyclable(Page page)
  */
 void
 _bt_delitems(Relation rel, Buffer buf,
-			 OffsetNumber *itemnos, int nitems)
+			 OffsetNumber *itemnos, int nitems,
+			 bool inVacuum)
 {
 	Page		page = BufferGetPage(buf);
 	BTPageOpaque opaque;
@@ -668,11 +669,12 @@ _bt_delitems(Relation rel, Buffer buf,
 	PageIndexMultiDelete(page, itemnos, nitems);
 
 	/*
-	 * We can clear the vacuum cycle ID since this page has certainly been
-	 * processed by the current vacuum scan.
+	 * If this is within VACUUM, we can clear the vacuum cycle ID since this
+	 * page has certainly been processed by the current vacuum scan.
 	 */
 	opaque = (BTPageOpaque) PageGetSpecialPointer(page);
-	opaque->btpo_cycleid = 0;
+	if (inVacuum)
+		opaque->btpo_cycleid = 0;
 
 	/*
 	 * Mark the page as not containing any LP_DEAD items.  This is not
diff --git a/src/backend/access/nbtree/nbtree.c b/src/backend/access/nbtree/nbtree.c
@@ -857,7 +857,7 @@ btvacuumpage(BTVacState *vstate, BlockNumber blkno, BlockNumber orig_blkno)
 		 */
 		if (ndeletable > 0)
 		{
-			_bt_delitems(rel, buf, deletable, ndeletable);
+			_bt_delitems(rel, buf, deletable, ndeletable, true);
 			stats->tuples_removed += ndeletable;
 			/* must recompute maxoff */
 			maxoff = PageGetMaxOffsetNumber(page);
diff --git a/src/include/access/nbtree.h b/src/include/access/nbtree.h
@@ -537,7 +537,7 @@ extern void _bt_relbuf(Relation rel, Buffer buf);
 extern void _bt_pageinit(Page page, Size size);
 extern bool _bt_page_recyclable(Page page);
 extern void _bt_delitems(Relation rel, Buffer buf,
-			 OffsetNumber *itemnos, int nitems);
+			 OffsetNumber *itemnos, int nitems, bool inVacuum);
 extern int _bt_pagedel(Relation rel, Buffer buf,
 			BTStack stack, bool vacuum_full);
 

Original file line number	Diff line number	Diff line change
`@@ -1978,7 +1978,7 @@ _bt_vacuum_one_page(Relation rel, Buffer buffer)`
`1978`	`1978`	`}`
`1979`	`1979`
`1980`	`1980`	`if (ndeletable > 0)`
`1981`		`- _bt_delitems(rel, buffer, deletable, ndeletable);`
	`1981`	`+ _bt_delitems(rel, buffer, deletable, ndeletable, false);`
`1982`	`1982`
`1983`	`1983`	`/*`
`1984`	`1984`	`* Note: if we didn't find any LP_DEAD items, then the page's`
Original file line number	Diff line number	Diff line change
`@@ -857,7 +857,7 @@ btvacuumpage(BTVacState *vstate, BlockNumber blkno, BlockNumber orig_blkno)`
`857`	`857`	`*/`
`858`	`858`	`if (ndeletable > 0)`
`859`	`859`	`{`
`860`		`- _bt_delitems(rel, buf, deletable, ndeletable);`
	`860`	`+ _bt_delitems(rel, buf, deletable, ndeletable, true);`
`861`	`861`	`stats->tuples_removed += ndeletable;`
`862`	`862`	`/* must recompute maxoff */`
`863`	`863`	`maxoff = PageGetMaxOffsetNumber(page);`