update relpages and reltuples when vacuum full (#7040)

* Update relation's stats in pg_class during vacuum full.

Hash index depends on estimation of numbers of tuples and pages of relations,
incorrect value could be a reason of significantly growing of index. Vacuum
full recreates heap and reindex all indexes before renewal stats. The patch
fixes that, so indexes will see correct values.

Backpatch to v10 only because earlier versions haven't usable hash index and
growing of hash index is a single user-visible symptom.

Author: Amit Kapila
Reviewed-by: Ashutosh Sharma, me
Discussion: https://www.postgresql.org/message-id/flat/20171115232922.5tomkxnw3iq6jsg7@inml.weebeastie.net

* Collect QE's relpages and reltuples to QD

And logic in swap_relation_files() to collect QE's relpages and reltuples to QD when doing vacuum full
Co-authored-by: NJimmy Yih <jyih@pivotal.io>
Co-authored-by: NAshwin Agrawal <aagrawal@pivotal.io>

* Add test

Add test to verify that relpages and reltuples has become proper numbers when vacuum full
Co-authored-by: NTaylor Vesely <tvesely@pivotal.io>

* Update PR pipeline failures
Reviewed-by: NAdam Berlin <aberlin@pivotal.io>
Reviewed-by: NAlexandra Wang <lewang@pivotal.io>
Reviewed-by: NJimmy Yih <jyih@pivotal.io>
Reviewed-by: NAshwin Agrawal <aagrawal@pivotal.io>
Reviewed-by: NTaylor Vesely <tvesely@pivotal.io>

src/backend/commands/cluster.c

@@ -65,6 +65,7 @@
 #include "cdb/cdbvars.h"
 #include "cdb/cdbdisp_query.h"
 #include "cdb/cdboidsync.h"
+#include "libpq/pqformat.h"
 
 /*
  * This struct is used to pass around the information on tables to be
@@ -839,6 +840,9 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex, bool verbose,
 	Relation	NewHeap,
 				OldHeap,
 				OldIndex;
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class	relform;
 	TupleDesc	oldTupDesc;
 	TupleDesc	newTupDesc;
 	int			natts;
@@ -857,6 +861,7 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex, bool verbose,
 	double		num_tuples = 0,
 				tups_vacuumed = 0,
 				tups_recently_dead = 0;
+	BlockNumber num_pages;
 	int			elevel = verbose ? INFO : DEBUG2;
 	PGRUsage	ru0;
 
@@ -1184,6 +1189,8 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex, bool verbose,
 	/* Reset rd_toastoid just to be tidy --- it shouldn't be looked at again */
 	NewHeap->rd_toastoid = InvalidOid;
 
+	num_pages = RelationGetNumberOfBlocks(NewHeap);
+
 	/* Log what we did */
 	ereport(elevel,
 			(errmsg("\"%s\": found %.0f removable, %.0f nonremovable row versions in %u pages",
@@ -1203,6 +1210,35 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex, bool verbose,
 		index_close(OldIndex, NoLock);
 	heap_close(OldHeap, NoLock);
 	heap_close(NewHeap, NoLock);
+
+	/* Update pg_class to reflect the correct values of pages and tuples. */
+	relRelation = heap_open(RelationRelationId, RowExclusiveLock);
+
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDNewHeap));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", OIDNewHeap);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+
+	relform->relpages = num_pages;
+	relform->reltuples = num_tuples;
+
+	/* Don't update the stats for pg_class.  See swap_relation_files. */
+	if (OIDOldHeap != RelationRelationId)
+	{
+		simple_heap_update(relRelation, &reltup->t_self, reltup);
+
+		/* keep the catalog indexes up to date */
+		CatalogUpdateIndexes(relRelation, reltup);
+	}
+	else
+		CacheInvalidateRelcacheByTuple(reltup);
+
+	/* Clean up. */
+	heap_freetuple(reltup);
+	heap_close(relRelation, RowExclusiveLock);
+
+	/* Make the update visible */
+	CommandCounterIncrement();
 }
 
 /*
@@ -1291,7 +1327,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					MultiXactId cutoffMulti,
 					Oid *mapped_tables)
 {
-	Relation	relRelation;
+	Relation	relRelation,
+				rel;
 	HeapTuple	reltup1,
 				reltup2;
 	Form_pg_class relform1,
@@ -1586,6 +1623,20 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							mapped_tables);
 	}
 
+	/* swap size statistics too, since new rel has freshly-updated stats */
+	if (swap_stats)
+	{
+		rel = relation_open(r1, AccessShareLock);
+
+		vac_update_relstats(rel, relform1->relpages, relform1->reltuples,
+							relform1->relallvisible,
+							relform1->relhaspkey,
+							relform1->relfrozenxid,
+							relform1->relminmxid,
+							false,
+							true /* isvacuum */);
+		relation_close(rel, AccessShareLock);
+	}
 	/* Clean up. */
 	heap_freetuple(reltup1);
 	heap_freetuple(reltup2);

src/backend/commands/vacuum.c

@@ -2423,6 +2423,8 @@ vacuum_rel(Relation onerel, Oid relid, VacuumStmt *vacstmt, LOCKMODE lmode,
 								   save_userid,
 								   save_sec_context | SECURITY_RESTRICTED_OPERATION);
 			dispatchVacuum(vacstmt, &stats_context);
+
+			vac_update_relstats_from_list(stats_context.updated_stats);
 		}
 	}
 	else

src/test/regress/expected/subselect_gp2.out

@@ -41,19 +41,19 @@ and usename='xxx' and datname='xxx';
                         QUERY PLAN                         
 -----------------------------------------------------------
  Hash Join
-   Hash Cond: (s.datid = d.oid)
+   Hash Cond: (s.usesysid = u.oid)
    InitPlan 1 (returns $0)
      ->  Result
    ->  Hash Join
-         Hash Cond: (u.oid = s.usesysid)
-         ->  Seq Scan on pg_authid u
-               Filter: (rolname = 'xxx'::name)
+         Hash Cond: (s.datid = d.oid)
+         ->  Function Scan on pg_stat_get_activity s
+               Filter: (query = $0)
          ->  Hash
-               ->  Function Scan on pg_stat_get_activity s
-                     Filter: (query = $0)
+               ->  Seq Scan on pg_database d
+                     Filter: (datname = 'xxx'::name)
    ->  Hash
-         ->  Seq Scan on pg_database d
-               Filter: (datname = 'xxx'::name)
+         ->  Seq Scan on pg_authid u
+               Filter: (rolname = 'xxx'::name)
  Optimizer: legacy query optimizer
 (15 rows)
 

src/test/regress/expected/vacuum_full_heap.out

@@ -17,7 +17,24 @@ select pg_relation_size('ivfheap') from gp_dist_random('gp_id') where gp_segment
 (1 row)
 
 -- show pages are truncated
+-- GPDB-specific: VACUUM FULL on heap gives proper relpages and reltuples
+select relname, relpages, reltuples, gp_segment_id from gp_dist_random('pg_class') where oid = 'vfheap'::regclass;
+ relname | relpages | reltuples | gp_segment_id 
+---------+----------+-----------+---------------
+ vfheap  |        0 |         0 |             0
+ vfheap  |        4 |       100 |             1
+ vfheap  |        0 |         0 |             2
+(3 rows)
+
 vacuum full vfheap;
+select relname, relpages, reltuples, gp_segment_id from gp_dist_random('pg_class') where oid = 'vfheap'::regclass;
+ relname | relpages | reltuples | gp_segment_id 
+---------+----------+-----------+---------------
+ vfheap  |        2 |        50 |             1
+ vfheap  |        0 |         0 |             2
+ vfheap  |        0 |         0 |             0
+(3 rows)
+
 select pg_relation_size('vfheap') from gp_dist_random('gp_id') where gp_segment_id = 1;
  pg_relation_size 
 ------------------

src/test/regress/sql/vacuum_full_heap.sql

@@ -9,7 +9,10 @@ select pg_relation_size('vfheap') from gp_dist_random('gp_id') where gp_segment_
 select pg_relation_size('ivfheap') from gp_dist_random('gp_id') where gp_segment_id = 1;
 
 -- show pages are truncated
+-- GPDB-specific: VACUUM FULL on heap gives proper relpages and reltuples
+select relname, relpages, reltuples, gp_segment_id from gp_dist_random('pg_class') where oid = 'vfheap'::regclass;
 vacuum full vfheap;
+select relname, relpages, reltuples, gp_segment_id from gp_dist_random('pg_class') where oid = 'vfheap'::regclass;
 select pg_relation_size('vfheap') from gp_dist_random('gp_id') where gp_segment_id = 1;
 select pg_relation_size('ivfheap') from gp_dist_random('gp_id') where gp_segment_id = 1;