Cleanup of new b-tree page deletion code.

When marking a branch as half-dead, a pointer to the top of the branch is
stored in the leaf block's hi-key. During normal operation, the high key
was left in place, and the block number was just stored in the ctid field
of the high key tuple, but in WAL replay, the high key was recreated as a
truncated tuple with zero columns. For the sake of easier debugging, also
truncate the tuple in normal operation, so that the page is identical
after WAL replay. Also, rename the 'downlink' field in the WAL record to
'topparent', as that seems like a more descriptive name. And make sure
it's set to invalid when unlinking the leaf page.

diff --git a/src/backend/access/nbtree/nbtpage.c b/src/backend/access/nbtree/nbtpage.c
index b95c1b886d4df4bd28df74848701f28df3b06ac8..c0ebb95ba8a3e56064a1a372a02a6c45ab20bdbc 100644 (file)
--- a/src/backend/access/nbtree/nbtpage.c
+++ b/src/backend/access/nbtree/nbtpage.c
@@ -1303,6 +1303,10 @@ _bt_pagedel(Relation rel, Buffer buf)
        return ndeleted;
 }
 
+/*
+ * First stage of page deletion.  Remove the downlink to the top of the
+ * branch being deleted, and mark the leaf page as half-dead.
+ */
 static bool
 _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
 {
@@ -1317,6 +1321,7 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
        OffsetNumber topoff;
        OffsetNumber nextoffset;
        IndexTuple      itup;
+       IndexTupleData trunctuple;
 
        page = BufferGetPage(leafbuf);
        opaque = (BTPageOpaque) PageGetSpecialPointer(page);
@@ -1406,12 +1411,17 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
        opaque = (BTPageOpaque) PageGetSpecialPointer(page);
        opaque->btpo_flags |= BTP_HALF_DEAD;
 
-       itemid = PageGetItemId(page, P_HIKEY);
-       itup = (IndexTuple) PageGetItem(page, itemid);
-       if (target == leafblkno)
-               ItemPointerSetInvalid(&(itup->t_tid));
+       PageIndexTupleDelete(page, P_HIKEY);
+       Assert(PageGetMaxOffsetNumber(page) == 0);
+       MemSet(&trunctuple, 0, sizeof(IndexTupleData));
+       trunctuple.t_info = sizeof(IndexTupleData);
+       if (target != leafblkno)
+               ItemPointerSet(&trunctuple.t_tid, target, P_HIKEY);
        else
-               ItemPointerSet(&(itup->t_tid), target, P_HIKEY);
+               ItemPointerSetInvalid(&trunctuple.t_tid);
+       if (PageAddItem(page, (Item) &trunctuple, sizeof(IndexTupleData), P_HIKEY,
+                                       false, false) == InvalidOffsetNumber)
+               elog(ERROR, "could not add dummy high key to half-dead page");
 
        /* Must mark buffers dirty before XLogInsert */
        MarkBufferDirty(topparent);
@@ -1427,7 +1437,10 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
                xlrec.target.node = rel->rd_node;
                ItemPointerSet(&(xlrec.target.tid), BufferGetBlockNumber(topparent), topoff);
                xlrec.leafblk = leafblkno;
-               xlrec.downlink = target;
+               if (target != leafblkno)
+                       xlrec.topparent = target;
+               else
+                       xlrec.topparent = InvalidBlockNumber;
 
                page = BufferGetPage(leafbuf);
                opaque = (BTPageOpaque) PageGetSpecialPointer(page);
@@ -1768,7 +1781,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, bool *rightsib_empty)
                xlrec.leafblk = leafblkno;
                xlrec.leafleftsib = leafleftsib;
                xlrec.leafrightsib = leafrightsib;
-               xlrec.downlink = nextchild;
+               xlrec.topparent = nextchild;
 
                rdata[0].data = (char *) &xlrec;
                rdata[0].len = SizeOfBtreeUnlinkPage;

diff --git a/src/backend/access/nbtree/nbtxlog.c b/src/backend/access/nbtree/nbtxlog.c
index ada2d6691f3a40a8b166c82954da87564fd53159..dc73f2a9cd54445c3f2d10d4aea09271e73ba390 100644 (file)
--- a/src/backend/access/nbtree/nbtxlog.c
+++ b/src/backend/access/nbtree/nbtxlog.c
@@ -870,8 +870,8 @@ btree_xlog_mark_page_halfdead(uint8 info, XLogRecPtr lsn, XLogRecord *record)
         */
        MemSet(&trunctuple, 0, sizeof(IndexTupleData));
        trunctuple.t_info = sizeof(IndexTupleData);
-       if (xlrec->downlink != InvalidBlockNumber)
-               ItemPointerSet(&trunctuple.t_tid, xlrec->downlink, P_HIKEY);
+       if (xlrec->topparent != InvalidBlockNumber)
+               ItemPointerSet(&trunctuple.t_tid, xlrec->topparent, P_HIKEY);
        else
                ItemPointerSetInvalid(&trunctuple.t_tid);
        if (PageAddItem(page, (Item) &trunctuple, sizeof(IndexTupleData), P_HIKEY,
@@ -1006,8 +1006,8 @@ btree_xlog_unlink_page(uint8 info, XLogRecPtr lsn, XLogRecord *record)
                /* Add a dummy hikey item */
                MemSet(&trunctuple, 0, sizeof(IndexTupleData));
                trunctuple.t_info = sizeof(IndexTupleData);
-               if (xlrec->downlink != InvalidBlockNumber)
-                       ItemPointerSet(&trunctuple.t_tid, xlrec->downlink, P_HIKEY);
+               if (xlrec->topparent != InvalidBlockNumber)
+                       ItemPointerSet(&trunctuple.t_tid, xlrec->topparent, P_HIKEY);
                else
                        ItemPointerSetInvalid(&trunctuple.t_tid);
                if (PageAddItem(page, (Item) &trunctuple, sizeof(IndexTupleData), P_HIKEY,

diff --git a/src/backend/access/rmgrdesc/nbtdesc.c b/src/backend/access/rmgrdesc/nbtdesc.c
index 89a91a20be199b6310bedd60f478fd504772aaaa..af7663b8cacfaa1a3ef582220aa4a18fd043ab4e 100644 (file)
--- a/src/backend/access/rmgrdesc/nbtdesc.c
+++ b/src/backend/access/rmgrdesc/nbtdesc.c
@@ -130,8 +130,8 @@ btree_desc(StringInfo buf, uint8 xl_info, char *rec)
 
                                appendStringInfoString(buf, "mark_page_halfdead: ");
                                out_target(buf, &(xlrec->target));
-                               appendStringInfo(buf, "; downlink %u; leaf %u; left %u; right %u",
-                                                                xlrec->downlink, xlrec->leafblk, xlrec->leftblk, xlrec->rightblk);
+                               appendStringInfo(buf, "; topparent %u; leaf %u; left %u; right %u",
+                                                                xlrec->topparent, xlrec->leafblk, xlrec->leftblk, xlrec->rightblk);
                                break;
                        }
                case XLOG_BTREE_UNLINK_PAGE_META:
@@ -143,8 +143,8 @@ btree_desc(StringInfo buf, uint8 xl_info, char *rec)
                                                                 xlrec->node.spcNode, xlrec->node.dbNode, xlrec->node.relNode);
                                appendStringInfo(buf, "dead %u; left %u; right %u; btpo_xact %u; ",
                                                                 xlrec->deadblk, xlrec->leftsib, xlrec->rightsib, xlrec->btpo_xact);
-                               appendStringInfo(buf, "leaf %u; leafleft %u; leafright %u; downlink %u",
-                                                                xlrec->leafblk, xlrec->leafleftsib, xlrec->leafrightsib, xlrec->downlink);
+                               appendStringInfo(buf, "leaf %u; leafleft %u; leafright %u; topparent %u",
+                                                                xlrec->leafblk, xlrec->leafleftsib, xlrec->leafrightsib, xlrec->topparent);
                                break;
                        }
                case XLOG_BTREE_NEWROOT:

diff --git a/src/include/access/nbtree.h b/src/include/access/nbtree.h
index 64c6982f50e69e943b3f27daabdf55283018fcf4..1a8b16d45e267207d5f9777c793d9fb785305523 100644 (file)
--- a/src/include/access/nbtree.h
+++ b/src/include/access/nbtree.h
@@ -379,13 +379,15 @@ typedef struct xl_btree_vacuum
 typedef struct xl_btree_mark_page_halfdead
 {
        xl_btreetid target;                     /* deleted tuple id in parent page */
+
+       /* information needed to recreate the leaf page: */
        BlockNumber leafblk;            /* leaf block ultimately being deleted */
        BlockNumber leftblk;            /* leaf block's left sibling, if any */
        BlockNumber rightblk;           /* leaf block's right sibling */
-       BlockNumber downlink;           /* next child down in the branch */
+       BlockNumber topparent;          /* topmost internal page in the branch */
 } xl_btree_mark_page_halfdead;
 
-#define SizeOfBtreeMarkPageHalfDead    (offsetof(xl_btree_mark_page_halfdead, downlink) + sizeof(BlockNumber))
+#define SizeOfBtreeMarkPageHalfDead    (offsetof(xl_btree_mark_page_halfdead, topparent) + sizeof(BlockNumber))
 
 /*
  * This is what we need to know about deletion of a btree page.  Note we do
@@ -406,7 +408,7 @@ typedef struct xl_btree_unlink_page
        BlockNumber     leafblk;
        BlockNumber     leafleftsib;
        BlockNumber     leafrightsib;
-       BlockNumber     downlink;               /* next child down in the branch */
+       BlockNumber     topparent;              /* next child down in the branch */
 
        TransactionId btpo_xact;        /* value of btpo.xact for use in recovery */
        /* xl_btree_metadata FOLLOWS IF XLOG_BTREE_UNLINK_PAGE_META */