Make VACUUM avoid waiting for a cleanup lock, where possible.

In a regular VACUUM, it's OK to skip pages for which a cleanup lock
isn't immediately available; the next VACUUM will deal with them.  If
we're scanning the entire relation to advance relfrozenxid, we might
need to wait, but only if there are tuples on the page that actually
require freezing.  These changes should greatly reduce the incidence
of of vacuum processes getting "stuck".

Simon Riggs and Robert Haas

Author: robertmhaas

src/backend/access/heap/heapam.c

@@ -3841,6 +3841,44 @@ heap_freeze_tuple(HeapTupleHeader tuple, TransactionId cutoff_xid,
 	return changed;
 }
 
+/*
+ * heap_tuple_needs_freeze
+ *
+ * Check to see whether any of the XID fields of a tuple (xmin, xmax, xvac)
+ * are older than the specified cutoff XID.  If so, return TRUE.
+ *
+ * It doesn't matter whether the tuple is alive or dead, we are checking
+ * to see if a tuple needs to be removed or frozen to avoid wraparound.
+ */
+bool
+heap_tuple_needs_freeze(HeapTupleHeader tuple, TransactionId cutoff_xid,
+				  Buffer buf)
+{
+	TransactionId xid;
+
+	xid = HeapTupleHeaderGetXmin(tuple);
+	if (TransactionIdIsNormal(xid) &&
+		TransactionIdPrecedes(xid, cutoff_xid))
+		return true;
+
+	if (!(tuple->t_infomask & HEAP_XMAX_IS_MULTI))
+	{
+		xid = HeapTupleHeaderGetXmax(tuple);
+		if (TransactionIdIsNormal(xid) &&
+			TransactionIdPrecedes(xid, cutoff_xid))
+			return true;
+	}
+
+	if (tuple->t_infomask & HEAP_MOVED)
+	{
+		xid = HeapTupleHeaderGetXvac(tuple);
+		if (TransactionIdIsNormal(xid) &&
+			TransactionIdPrecedes(xid, cutoff_xid))
+			return true;
+	}
+
+	return false;
+}
 
 /* ----------------
  *		heap_markpos	- mark scan position

src/backend/commands/vacuumlazy.c

@@ -117,6 +117,7 @@ static BufferAccessStrategy vac_strategy;
 static void lazy_scan_heap(Relation onerel, LVRelStats *vacrelstats,
 			   Relation *Irel, int nindexes, bool scan_all);
 static void lazy_vacuum_heap(Relation onerel, LVRelStats *vacrelstats);
+static bool lazy_check_needs_freeze(Buffer buf);
 static void lazy_vacuum_index(Relation indrel,
 				  IndexBulkDeleteResult **stats,
 				  LVRelStats *vacrelstats);
@@ -453,8 +454,6 @@ lazy_scan_heap(Relation onerel, LVRelStats *vacrelstats,
 
 		vacuum_delay_point();
 
-		vacrelstats->scanned_pages++;
-
 		/*
 		 * If we are close to overrunning the available space for dead-tuple
 		 * TIDs, pause and do a cycle of vacuuming before we tackle this page.
@@ -486,7 +485,41 @@ lazy_scan_heap(Relation onerel, LVRelStats *vacrelstats,
 								 RBM_NORMAL, vac_strategy);
 
 		/* We need buffer cleanup lock so that we can prune HOT chains. */
-		LockBufferForCleanup(buf);
+		if (!ConditionalLockBufferForCleanup(buf))
+		{
+			/*
+			 * It's OK to skip vacuuming a page, as long as its not got data
+			 * that needs to be cleaned for wraparound avoidance.
+			 */
+			if (!scan_all)
+			{
+				ReleaseBuffer(buf);
+				continue;
+			}
+
+			/*
+			 * If this is a wraparound checking vacuum, then we read the page
+			 * with share lock to see if any xids need to be frozen. If the
+			 * page doesn't need attention we just skip and continue. If it
+			 * does, we wait for cleanup lock.
+			 *
+			 * We could defer the lock request further by remembering the page
+			 * and coming back to it later, of we could even register
+			 * ourselves for multiple buffers and then service whichever one
+			 * is received first.  For now, this seems good enough.
+			 */
+			LockBuffer(buf, BUFFER_LOCK_SHARE);
+			if (!lazy_check_needs_freeze(buf))
+			{
+				UnlockReleaseBuffer(buf);
+				continue;
+			}
+			LockBuffer(buf, BUFFER_LOCK_UNLOCK);
+			LockBufferForCleanup(buf);
+			/* drop through to normal processing */
+		}
+
+		vacrelstats->scanned_pages++;
 
 		page = BufferGetPage(buf);
 
@@ -932,7 +965,8 @@ lazy_vacuum_heap(Relation onerel, LVRelStats *vacrelstats)
 		tblk = ItemPointerGetBlockNumber(&vacrelstats->dead_tuples[tupindex]);
 		buf = ReadBufferExtended(onerel, MAIN_FORKNUM, tblk, RBM_NORMAL,
 								 vac_strategy);
-		LockBufferForCleanup(buf);
+		if (!ConditionalLockBufferForCleanup(buf))
+			continue;
 		tupindex = lazy_vacuum_page(onerel, tblk, buf, tupindex, vacrelstats);
 
 		/* Now that we've compacted the page, record its available space */
@@ -1009,6 +1043,50 @@ lazy_vacuum_page(Relation onerel, BlockNumber blkno, Buffer buffer,
 	return tupindex;
 }
 
+/*
+ *	lazy_check_needs_freeze() -- scan page to see if any tuples
+ *					 need to be cleaned to avoid wraparound
+ *
+ * Returns true if the page needs to be vacuumed using cleanup lock.
+ */
+static bool
+lazy_check_needs_freeze(Buffer buf)
+{
+	Page		page;
+	OffsetNumber offnum,
+				maxoff;
+	HeapTupleHeader tupleheader;
+
+	page = BufferGetPage(buf);
+
+	if (PageIsNew(page) || PageIsEmpty(page))
+	{
+		/* PageIsNew probably shouldn't happen... */
+		return false;
+	}
+
+	maxoff = PageGetMaxOffsetNumber(page);
+	for (offnum = FirstOffsetNumber;
+		 offnum <= maxoff;
+		 offnum = OffsetNumberNext(offnum))
+	{
+		ItemId		itemid;
+
+		itemid = PageGetItemId(page, offnum);
+
+		if (!ItemIdIsNormal(itemid))
+			continue;
+
+		tupleheader = (HeapTupleHeader) PageGetItem(page, itemid);
+
+		if (heap_tuple_needs_freeze(tupleheader, FreezeLimit, buf))
+			return true;
+	}						/* scan along page */
+
+	return false;
+}
+
+
 /*
  *	lazy_vacuum_index() -- vacuum one index relation.
  *

src/include/access/heapam.h

@@ -111,6 +111,8 @@ extern HTSU_Result heap_lock_tuple(Relation relation, HeapTuple tuple,
 extern void heap_inplace_update(Relation relation, HeapTuple tuple);
 extern bool heap_freeze_tuple(HeapTupleHeader tuple, TransactionId cutoff_xid,
 				  Buffer buf);
+extern bool heap_tuple_needs_freeze(HeapTupleHeader tuple, TransactionId cutoff_xid,
+				  Buffer buf);
 
 extern Oid	simple_heap_insert(Relation relation, HeapTuple tup);
 extern void simple_heap_delete(Relation relation, ItemPointer tid);