Handle new HOT chains in index-build table scans

When a table is scanned by heapam_index_build_range_scan (née
IndexBuildHeapScan) and the table lock being held allows concurrent data
changes, it is possible for new HOT chains to sprout in a page that were
unknown when the scan of a page happened.  This leads to an error such
as
  ERROR:  failed to find parent tuple for heap-only tuple at (X,Y) in table "tbl"
because the root tuple was not present when we first obtained the list
of the page's root tuples.  This can be fixed by re-obtaining the list
of root tuples, if we see that a heap-only tuple appears to point to a
non-existing root.

This was reported by Anastasia as occurring for BRIN summarization
(which exists since 9.5), but I think it could theoretically also happen
with CREATE INDEX CONCURRENTLY (much older) or REINDEX CONCURRENTLY
(very recent).  It seems a happy coincidence that BRIN forces us to
backpatch this all the way to 9.5.

Reported-by: Anastasia Lubennikova 
Diagnosed-by: Anastasia Lubennikova 
Co-authored-by: Anastasia Lubennikova 
Co-authored-by: Álvaro Herrera 
Discussion: https://postgr.es/m/602d8487-f0b2-5486-0088-0f372b2549fa@postgrespro.ru
Backpatch: 9.5 - master

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 56b35622f1a4faa6bcd4d2de6b304fedee5482e0..9d28288649268c72ed4b76a4ba4be42c2f7f59a2 100644 (file)
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1322,6 +1322,12 @@ heapam_index_build_range_scan(Relation heapRelation,
         * buffer continuously while visiting the page, so no pruning
         * operation can occur either.
         *
+        * In cases with only ShareUpdateExclusiveLock on the table, it's
+        * possible for some HOT tuples to appear that we didn't know about
+        * when we first read the page.  To handle that case, we re-obtain the
+        * list of root offsets when a HOT tuple points to a root item that we
+        * don't know about.
+        *
         * Also, although our opinions about tuple liveness could change while
         * we scan the page (due to concurrent transaction commits/aborts),
         * the chain root locations won't, so this info doesn't need to be
@@ -1623,6 +1629,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
            offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+           /*
+            * If a HOT tuple points to a root that we don't know
+            * about, obtain root items afresh.  If that still fails,
+            * report it as corruption.
+            */
+           if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+           {
+               Page    page = BufferGetPage(hscan->rs_cbuf);
+
+               LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+               heap_get_root_tuples(page, root_offsets);
+               LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+           }
+
            if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
                ereport(ERROR,
                        (errcode(ERRCODE_DATA_CORRUPTED),

diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 1794cfd8d9aab2c4cd4bb25a587a69e7b376f4d4..7740e1b93952f66d395d7536b6b4e5c00a3f28eb 100644 (file)
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
    OffsetNumber offnum,
                maxoff;
 
-   MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+   MemSet(root_offsets, InvalidOffsetNumber,
+          MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
    maxoff = PageGetMaxOffsetNumber(page);
    for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))