3 files changed, 40 insertions, 6 deletions

diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
index 201d1f5a1b4..fa415ab06f9 100644
--- a/src/backend/access/heap/heapam.c
+++ b/src/backend/access/heap/heapam.c
@@ -2308,6 +2308,7 @@ heap_get_latest_tid(Relation relation,
 		 */
 		if ((tp.t_data->t_infomask & HEAP_XMAX_INVALID) ||
 			HeapTupleHeaderIsOnlyLocked(tp.t_data) ||
+			HeapTupleHeaderIndicatesMovedPartitions(tp.t_data) ||
 			ItemPointerEquals(&tp.t_self, &tp.t_data->t_ctid))
 		{
 			UnlockReleaseBuffer(buffer);
@@ -3041,6 +3042,8 @@ xmax_infomask_changed(uint16 new_infomask, uint16 old_infomask)
  *	crosscheck - if not InvalidSnapshot, also check tuple against this
  *	wait - true if should wait for any conflicting update to commit/abort
  *	hufd - output parameter, filled in failure cases (see below)
+ *	changingPart - true iff the tuple is being moved to another partition
+ *		table due to an update of the partition key. Otherwise, false.
  *
  * Normal, successful return value is HeapTupleMayBeUpdated, which
  * actually means we did delete it.  Failure return codes are
@@ -3056,7 +3059,7 @@ xmax_infomask_changed(uint16 new_infomask, uint16 old_infomask)
 HTSU_Result
 heap_delete(Relation relation, ItemPointer tid,
 			CommandId cid, Snapshot crosscheck, bool wait,
-			HeapUpdateFailureData *hufd)
+			HeapUpdateFailureData *hufd, bool changingPart)
 {
 	HTSU_Result result;
 	TransactionId xid = GetCurrentTransactionId();
@@ -3325,6 +3328,10 @@ l1:
 	/* Make sure there is no forward chain link in t_ctid */
 	tp.t_data->t_ctid = tp.t_self;
 
+	/* Signal that this is actually a move into another partition */
+	if (changingPart)
+		HeapTupleHeaderSetMovedPartitions(tp.t_data);
+
 	MarkBufferDirty(buffer);
 
 	/*
@@ -3342,7 +3349,11 @@ l1:
 		if (RelationIsAccessibleInLogicalDecoding(relation))
 			log_heap_new_cid(relation, &tp);
 
-		xlrec.flags = all_visible_cleared ? XLH_DELETE_ALL_VISIBLE_CLEARED : 0;
+		xlrec.flags = 0;
+		if (all_visible_cleared)
+			xlrec.flags |= XLH_DELETE_ALL_VISIBLE_CLEARED;
+		if (changingPart)
+			xlrec.flags |= XLH_DELETE_IS_PARTITION_MOVE;
 		xlrec.infobits_set = compute_infobits(tp.t_data->t_infomask,
 											  tp.t_data->t_infomask2);
 		xlrec.offnum = ItemPointerGetOffsetNumber(&tp.t_self);
@@ -3450,7 +3461,7 @@ simple_heap_delete(Relation relation, ItemPointer tid)
 	result = heap_delete(relation, tid,
 						 GetCurrentCommandId(true), InvalidSnapshot,
 						 true /* wait for commit */ ,
-						 &hufd);
+						 &hufd, false /* changingPart */);
 	switch (result)
 	{
 		case HeapTupleSelfUpdated:
@@ -6051,6 +6062,7 @@ l4:
 next:
 		/* if we find the end of update chain, we're done. */
 		if (mytup.t_data->t_infomask & HEAP_XMAX_INVALID ||
+			HeapTupleHeaderIndicatesMovedPartitions(mytup.t_data) ||
 			ItemPointerEquals(&mytup.t_self, &mytup.t_data->t_ctid) ||
 			HeapTupleHeaderIsOnlyLocked(mytup.t_data))
 		{
@@ -6102,7 +6114,12 @@ static HTSU_Result
 heap_lock_updated_tuple(Relation rel, HeapTuple tuple, ItemPointer ctid,
 						TransactionId xid, LockTupleMode mode)
 {
-	if (!ItemPointerEquals(&tuple->t_self, ctid))
+	/*
+	 * If the tuple has not been updated, or has moved into another partition
+	 * (effectively a delete) stop here.
+	 */
+	if (!HeapTupleHeaderIndicatesMovedPartitions(tuple->t_data) &&
+		!ItemPointerEquals(&tuple->t_self, ctid))
 	{
 		/*
 		 * If this is the first possibly-multixact-able operation in the
@@ -8493,8 +8510,11 @@ heap_xlog_delete(XLogReaderState *record)
 		if (xlrec->flags & XLH_DELETE_ALL_VISIBLE_CLEARED)
 			PageClearAllVisible(page);
 
-		/* Make sure there is no forward chain link in t_ctid */
-		htup->t_ctid = target_tid;
+		/* Make sure t_ctid is set correctly */
+		if (xlrec->flags & XLH_DELETE_IS_PARTITION_MOVE)
+			HeapTupleHeaderSetMovedPartitions(htup);
+		else
+			htup->t_ctid = target_tid;
 		PageSetLSN(page, lsn);
 		MarkBufferDirty(buffer);
 	}
@@ -9422,6 +9442,13 @@ heap_mask(char *pagedata, BlockNumber blkno)
 			 */
 			if (HeapTupleHeaderIsSpeculative(page_htup))
 				ItemPointerSet(&page_htup->t_ctid, blkno, off);
+
+			/*
+			 * NB: Not ignoring ctid changes due to the tuple having moved
+			 * (i.e. HeapTupleHeaderIndicatesMovedPartitions), because that's
+			 * important information that needs to be in-sync between primary
+			 * and standby, and thus is WAL logged.
+			 */
 		}
 
 		/*
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index f67d7d15df1..c2f5343dac8 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -552,6 +552,9 @@ heap_prune_chain(Relation relation, Buffer buffer, OffsetNumber rootoffnum,
 		if (!HeapTupleHeaderIsHotUpdated(htup))
 			break;
 
+		/* HOT implies it can't have moved to different partition */
+		Assert(!HeapTupleHeaderIndicatesMovedPartitions(htup));
+
 		/*
 		 * Advance to next chain member.
 		 */
@@ -823,6 +826,9 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 			if (!HeapTupleHeaderIsHotUpdated(htup))
 				break;
 
+			/* HOT implies it can't have moved to different partition */
+			Assert(!HeapTupleHeaderIndicatesMovedPartitions(htup));
+
 			nextoffnum = ItemPointerGetOffsetNumber(&htup->t_ctid);
 			priorXmax = HeapTupleHeaderGetUpdateXid(htup);
 		}
diff --git a/src/backend/access/heap/rewriteheap.c b/src/backend/access/heap/rewriteheap.c
index 7d466c2588c..8d3c861a330 100644
--- a/src/backend/access/heap/rewriteheap.c
+++ b/src/backend/access/heap/rewriteheap.c
@@ -424,6 +424,7 @@ rewrite_heap_tuple(RewriteState state,
 	 */
 	if (!((old_tuple->t_data->t_infomask & HEAP_XMAX_INVALID) ||
 		  HeapTupleHeaderIsOnlyLocked(old_tuple->t_data)) &&
+		!HeapTupleHeaderIndicatesMovedPartitions(old_tuple->t_data) &&
 		!(ItemPointerEquals(&(old_tuple->t_self),
 							&(old_tuple->t_data->t_ctid))))
 	{