Improve concurrency of foreign key locking

This patch introduces two additional lock modes for tuples: "SELECT FOR
KEY SHARE" and "SELECT FOR NO KEY UPDATE".  These don't block each
other, in contrast with already existing "SELECT FOR SHARE" and "SELECT
FOR UPDATE".  UPDATE commands that do not modify the values stored in
the columns that are part of the key of the tuple now grab a SELECT FOR
NO KEY UPDATE lock on the tuple, allowing them to proceed concurrently
with tuple locks of the FOR KEY SHARE variety.

Foreign key triggers now use FOR KEY SHARE instead of FOR SHARE; this
means the concurrency improvement applies to them, which is the whole
point of this patch.

The added tuple lock semantics require some rejiggering of the multixact
module, so that the locking level that each transaction is holding can
be stored alongside its Xid.  Also, multixacts now need to persist
across server restarts and crashes, because they can now represent not
only tuple locks, but also tuple updates.  This means we need more
careful tracking of lifetime of pg_multixact SLRU files; since they now
persist longer, we require more infrastructure to figure out when they
can be removed.  pg_upgrade also needs to be careful to copy
pg_multixact files over from the old server to the new, or at least part
of multixact.c state, depending on the versions of the old and new
servers.

Tuple time qualification rules (HeapTupleSatisfies routines) need to be
careful not to consider tuples with the "is multi" infomask bit set as
being only locked; they might need to look up MultiXact values (i.e.
possibly do pg_multixact I/O) to find out the Xid that updated a tuple,
whereas they previously were assured to only use information readily
available from the tuple header.  This is considered acceptable, because
the extra I/O would involve cases that would previously cause some
commands to block waiting for concurrent transactions to finish.

Another important change is the fact that locking tuples that have
previously been updated causes the future versions to be marked as
locked, too; this is essential for correctness of foreign key checks.
This causes additional WAL-logging, also (there was previously a single
WAL record for a locked tuple; now there are as many as updated copies
of the tuple there exist.)

With all this in place, contention related to tuples being checked by
foreign key rules should be much reduced.

As a bonus, the old behavior that a subtransaction grabbing a stronger
tuple lock than the parent (sub)transaction held on a given tuple and
later aborting caused the weaker lock to be lost, has been fixed.

Many new spec files were added for isolation tester framework, to ensure
overall behavior is sane.  There's probably room for several more tests.

There were several reviewers of this patch; in particular, Noah Misch
and Andres Freund spent considerable time in it.  Original idea for the
patch came from Simon Riggs, after a problem report by Joel Jacobson.
Most code is from me, with contributions from Marti Raudsepp, Alexander
Shulgin, Noah Misch and Andres Freund.

This patch was discussed in several pgsql-hackers threads; the most
important start at the following message-ids:
	AANLkTimo9XVcEzfiBR-ut3KVNDkjm2Vxh+t8kAmWjPuv@mail.gmail.com
	1290721684-sup-3951@alvh.no-ip.org
	1294953201-sup-2099@alvh.no-ip.org
	1320343602-sup-2290@alvh.no-ip.org
	1339690386-sup-8927@alvh.no-ip.org
	4FE5FF020200002500048A3D@gw.wicourts.gov
	4FEAB90A0200002500048B7D@gw.wicourts.gov

1 files changed, 25 insertions, 12 deletions

diff --git a/src/backend/commands/cluster.c b/src/backend/commands/cluster.c
index 238781b6a70..c0cb2f66545 100644
--- a/src/backend/commands/cluster.c
+++ b/src/backend/commands/cluster.c
@@ -17,6 +17,7 @@
  */
 #include "postgres.h"
 
+#include "access/multixact.h"
 #include "access/relscan.h"
 #include "access/rewriteheap.h"
 #include "access/transam.h"
@@ -65,7 +66,8 @@ static void rebuild_relation(Relation OldHeap, Oid indexOid,
 				 int freeze_min_age, int freeze_table_age, bool verbose);
 static void copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex,
 			   int freeze_min_age, int freeze_table_age, bool verbose,
-			   bool *pSwapToastByContent, TransactionId *pFreezeXid);
+			   bool *pSwapToastByContent, TransactionId *pFreezeXid,
+			   MultiXactId *pFreezeMulti);
 static List *get_tables_to_cluster(MemoryContext cluster_context);
 static void reform_and_rewrite_tuple(HeapTuple tuple,
 						 TupleDesc oldTupDesc, TupleDesc newTupDesc,
@@ -549,6 +551,7 @@ rebuild_relation(Relation OldHeap, Oid indexOid,
 	bool		is_system_catalog;
 	bool		swap_toast_by_content;
 	TransactionId frozenXid;
+	MultiXactId	frozenMulti;
 
 	/* Mark the correct index as clustered */
 	if (OidIsValid(indexOid))
@@ -566,14 +569,14 @@ rebuild_relation(Relation OldHeap, Oid indexOid,
 	/* Copy the heap data into the new table in the desired order */
 	copy_heap_data(OIDNewHeap, tableOid, indexOid,
 				   freeze_min_age, freeze_table_age, verbose,
-				   &swap_toast_by_content, &frozenXid);
+				   &swap_toast_by_content, &frozenXid, &frozenMulti);
 
 	/*
 	 * Swap the physical files of the target and transient tables, then
 	 * rebuild the target's indexes and throw away the transient table.
 	 */
 	finish_heap_swap(tableOid, OIDNewHeap, is_system_catalog,
-					 swap_toast_by_content, false, frozenXid);
+					 swap_toast_by_content, false, frozenXid, frozenMulti);
 }
 
 
@@ -706,7 +709,8 @@ make_new_heap(Oid OIDOldHeap, Oid NewTableSpace)
 static void
 copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex,
 			   int freeze_min_age, int freeze_table_age, bool verbose,
-			   bool *pSwapToastByContent, TransactionId *pFreezeXid)
+			   bool *pSwapToastByContent, TransactionId *pFreezeXid,
+			   MultiXactId *pFreezeMulti)
 {
 	Relation	NewHeap,
 				OldHeap,
@@ -722,6 +726,7 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex,
 	bool		is_system_catalog;
 	TransactionId OldestXmin;
 	TransactionId FreezeXid;
+	MultiXactId	MultiXactFrzLimit;
 	RewriteState rwstate;
 	bool		use_sort;
 	Tuplesortstate *tuplesort;
@@ -822,7 +827,7 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex,
 	 */
 	vacuum_set_xid_limits(freeze_min_age, freeze_table_age,
 						  OldHeap->rd_rel->relisshared,
-						  &OldestXmin, &FreezeXid, NULL);
+						  &OldestXmin, &FreezeXid, NULL, &MultiXactFrzLimit);
 
 	/*
 	 * FreezeXid will become the table's new relfrozenxid, and that mustn't go
@@ -831,14 +836,16 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex,
 	if (TransactionIdPrecedes(FreezeXid, OldHeap->rd_rel->relfrozenxid))
 		FreezeXid = OldHeap->rd_rel->relfrozenxid;
 
-	/* return selected value to caller */
+	/* return selected values to caller */
 	*pFreezeXid = FreezeXid;
+	*pFreezeMulti = MultiXactFrzLimit;
 
 	/* Remember if it's a system catalog */
 	is_system_catalog = IsSystemRelation(OldHeap);
 
 	/* Initialize the rewrite operation */
-	rwstate = begin_heap_rewrite(NewHeap, OldestXmin, FreezeXid, use_wal);
+	rwstate = begin_heap_rewrite(NewHeap, OldestXmin, FreezeXid,
+								 MultiXactFrzLimit, use_wal);
 
 	/*
 	 * Decide whether to use an indexscan or seqscan-and-optional-sort to scan
@@ -966,9 +973,8 @@ copy_heap_data(Oid OIDNewHeap, Oid OIDOldHeap, Oid OIDOldIndex,
 				/*
 				 * Similar situation to INSERT_IN_PROGRESS case.
 				 */
-				Assert(!(tuple->t_data->t_infomask & HEAP_XMAX_IS_MULTI));
 				if (!is_system_catalog &&
-					!TransactionIdIsCurrentTransactionId(HeapTupleHeaderGetXmax(tuple->t_data)))
+					!TransactionIdIsCurrentTransactionId(HeapTupleHeaderGetUpdateXid(tuple->t_data)))
 					elog(WARNING, "concurrent delete in progress within table \"%s\"",
 						 RelationGetRelationName(OldHeap));
 				/* treat as recently dead */
@@ -1097,6 +1103,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					TransactionId frozenXid,
+					MultiXactId frozenMulti,
 					Oid *mapped_tables)
 {
 	Relation	relRelation;
@@ -1204,11 +1211,13 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	 * and then fail to commit the pg_class update.
 	 */
 
-	/* set rel1's frozen Xid */
+	/* set rel1's frozen Xid and minimum MultiXid */
 	if (relform1->relkind != RELKIND_INDEX)
 	{
 		Assert(TransactionIdIsNormal(frozenXid));
 		relform1->relfrozenxid = frozenXid;
+		Assert(MultiXactIdIsValid(frozenMulti));
+		relform1->relminmxid = frozenMulti;
 	}
 
 	/* swap size statistics too, since new rel has freshly-updated stats */
@@ -1272,6 +1281,7 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									frozenXid,
+									frozenMulti,
 									mapped_tables);
 			}
 			else
@@ -1361,6 +1371,7 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							InvalidTransactionId,
+							InvalidMultiXactId,
 							mapped_tables);
 
 	/* Clean up. */
@@ -1398,7 +1409,8 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 				 bool is_system_catalog,
 				 bool swap_toast_by_content,
 				 bool check_constraints,
-				 TransactionId frozenXid)
+				 TransactionId frozenXid,
+				 MultiXactId frozenMulti)
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
@@ -1414,7 +1426,8 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	 */
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
-						swap_toast_by_content, frozenXid, mapped_tables);
+						swap_toast_by_content, frozenXid, frozenMulti,
+						mapped_tables);
 
 	/*
 	 * If it's a system catalog, queue an sinval message to flush all