snapshot scalability: Don't compute global horizons while building snapshots.

To make GetSnapshotData() more scalable, it cannot not look at at each proc's
xmin: While snapshot contents do not need to change whenever a read-only
transaction commits or a snapshot is released, a proc's xmin is modified in
those cases. The frequency of xmin modifications leads to, particularly on
higher core count systems, many cache misses inside GetSnapshotData(), despite
the data underlying a snapshot not changing. That is the most
significant source of GetSnapshotData() scaling poorly on larger systems.

Without accessing xmins, GetSnapshotData() cannot calculate accurate horizons /
thresholds as it has so far. But we don't really have to: The horizons don't
actually change that much between GetSnapshotData() calls. Nor are the horizons
actually used every time a snapshot is built.

The trick this commit introduces is to delay computation of accurate horizons
until there use and using horizon boundaries to determine whether accurate
horizons need to be computed.

The use of RecentGlobal[Data]Xmin to decide whether a row version could be
removed has been replaces with new GlobalVisTest* functions.  These use two
thresholds to determine whether a row can be pruned:
1) definitely_needed, indicating that rows deleted by XIDs >= definitely_needed
   are definitely still visible.
2) maybe_needed, indicating that rows deleted by XIDs < maybe_needed can
   definitely be removed
GetSnapshotData() updates definitely_needed to be the xmin of the computed
snapshot.

When testing whether a row can be removed (with GlobalVisTestIsRemovableXid())
and the tested XID falls in between the two (i.e. XID >= maybe_needed && XID <
definitely_needed) the boundaries can be recomputed to be more accurate. As it
is not cheap to compute accurate boundaries, we limit the number of times that
happens in short succession.  As the boundaries used by
GlobalVisTestIsRemovableXid() are never reset (with maybe_needed updated by
GetSnapshotData()), it is likely that further test can benefit from an earlier
computation of accurate horizons.

To avoid regressing performance when old_snapshot_threshold is set (as that
requires an accurate horizon to be computed), heap_page_prune_opt() doesn't
unconditionally call TransactionIdLimitedForOldSnapshots() anymore. Both the
computation of the limited horizon, and the triggering of errors (with
SetOldSnapshotThresholdTimestamp()) is now only done when necessary to remove
tuples.

This commit just removes the accesses to PGXACT->xmin from
GetSnapshotData(), but other members of PGXACT residing in the same
cache line are accessed. Therefore this in itself does not result in a
significant improvement. Subsequent commits will take advantage of the
fact that GetSnapshotData() now does not need to access xmins anymore.

Note: This contains a workaround in heap_page_prune_opt() to keep the
snapshot_too_old tests working. While that workaround is ugly, the tests
currently are not meaningful, and it seems best to address them separately.

Author: Andres Freund <andres@anarazel.de>
Reviewed-By: Robert Haas <robertmhaas@gmail.com>
Reviewed-By: Thomas Munro <thomas.munro@gmail.com>
Reviewed-By: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/20200301083601.ews6hz5dduc3w2se@alap3.anarazel.de

2 files changed, 34 insertions, 9 deletions

diff --git a/src/include/utils/snapmgr.h b/src/include/utils/snapmgr.h
index ffb4ba3adfb..b6b403e2931 100644
--- a/src/include/utils/snapmgr.h
+++ b/src/include/utils/snapmgr.h
@@ -52,13 +52,12 @@ extern Size SnapMgrShmemSize(void);
 extern void SnapMgrInit(void);
 extern TimestampTz GetSnapshotCurrentTimestamp(void);
 extern TimestampTz GetOldSnapshotThresholdTimestamp(void);
+extern void SnapshotTooOldMagicForTest(void);
 
 extern bool FirstSnapshotSet;
 
 extern PGDLLIMPORT TransactionId TransactionXmin;
 extern PGDLLIMPORT TransactionId RecentXmin;
-extern PGDLLIMPORT TransactionId RecentGlobalXmin;
-extern PGDLLIMPORT TransactionId RecentGlobalDataXmin;
 
 /* Variables representing various special snapshot semantics */
 extern PGDLLIMPORT SnapshotData SnapshotSelfData;
@@ -78,11 +77,12 @@ extern PGDLLIMPORT SnapshotData CatalogSnapshotData;
 
 /*
  * Similarly, some initialization is required for a NonVacuumable snapshot.
- * The caller must supply the xmin horizon to use (e.g., RecentGlobalXmin).
+ * The caller must supply the visibility cutoff state to use (c.f.
+ * GlobalVisTestFor()).
  */
-#define InitNonVacuumableSnapshot(snapshotdata, xmin_horizon)  \
+#define InitNonVacuumableSnapshot(snapshotdata, vistestp)  \
 	((snapshotdata).snapshot_type = SNAPSHOT_NON_VACUUMABLE, \
-	 (snapshotdata).xmin = (xmin_horizon))
+	 (snapshotdata).vistest = (vistestp))
 
 /*
  * Similarly, some initialization is required for SnapshotToast.  We need
@@ -98,6 +98,11 @@ extern PGDLLIMPORT SnapshotData CatalogSnapshotData;
 	((snapshot)->snapshot_type == SNAPSHOT_MVCC || \
 	 (snapshot)->snapshot_type == SNAPSHOT_HISTORIC_MVCC)
 
+static inline bool
+OldSnapshotThresholdActive(void)
+{
+	return old_snapshot_threshold >= 0;
+}
 
 extern Snapshot GetTransactionSnapshot(void);
 extern Snapshot GetLatestSnapshot(void);
@@ -121,8 +126,6 @@ extern void UnregisterSnapshot(Snapshot snapshot);
 extern Snapshot RegisterSnapshotOnOwner(Snapshot snapshot, ResourceOwner owner);
 extern void UnregisterSnapshotFromOwner(Snapshot snapshot, ResourceOwner owner);
 
-extern FullTransactionId GetFullRecentGlobalXmin(void);
-
 extern void AtSubCommit_Snapshot(int level);
 extern void AtSubAbort_Snapshot(int level);
 extern void AtEOXact_Snapshot(bool isCommit, bool resetXmin);
@@ -131,14 +134,30 @@ extern void ImportSnapshot(const char *idstr);
 extern bool XactHasExportedSnapshots(void);
 extern void DeleteAllExportedSnapshotFiles(void);
 extern bool ThereAreNoPriorRegisteredSnapshots(void);
-extern TransactionId TransactionIdLimitedForOldSnapshots(TransactionId recentXmin,
-														 Relation relation);
+extern bool TransactionIdLimitedForOldSnapshots(TransactionId recentXmin,
+												Relation relation,
+												TransactionId *limit_xid,
+												TimestampTz *limit_ts);
+extern void SetOldSnapshotThresholdTimestamp(TimestampTz ts, TransactionId xlimit);
 extern void MaintainOldSnapshotTimeMapping(TimestampTz whenTaken,
 										   TransactionId xmin);
 
 extern char *ExportSnapshot(Snapshot snapshot);
 
 /*
+ * These live in procarray.c because they're intimately linked to the
+ * procarray contents, but thematically they better fit into snapmgr.h.
+ */
+typedef struct GlobalVisState GlobalVisState;
+extern GlobalVisState *GlobalVisTestFor(Relation rel);
+extern bool GlobalVisTestIsRemovableXid(GlobalVisState *state, TransactionId xid);
+extern bool GlobalVisTestIsRemovableFullXid(GlobalVisState *state, FullTransactionId fxid);
+extern FullTransactionId GlobalVisTestNonRemovableFullHorizon(GlobalVisState *state);
+extern TransactionId GlobalVisTestNonRemovableHorizon(GlobalVisState *state);
+extern bool GlobalVisCheckRemovableXid(Relation rel, TransactionId xid);
+extern bool GlobalVisIsRemovableFullXid(Relation rel, FullTransactionId fxid);
+
+/*
  * Utility functions for implementing visibility routines in table AMs.
  */
 extern bool XidInMVCCSnapshot(TransactionId xid, Snapshot snapshot);
diff --git a/src/include/utils/snapshot.h b/src/include/utils/snapshot.h
index 4796edb63aa..35b1f05bea6 100644
--- a/src/include/utils/snapshot.h
+++ b/src/include/utils/snapshot.h
@@ -193,6 +193,12 @@ typedef struct SnapshotData
 	uint32		speculativeToken;
 
 	/*
+	 * For SNAPSHOT_NON_VACUUMABLE (and hopefully more in the future) this is
+	 * used to determine whether row could be vacuumed.
+	 */
+	struct GlobalVisState *vistest;
+
+	/*
 	 * Book-keeping information, used by the snapshot manager
 	 */
 	uint32		active_count;	/* refcount on ActiveSnapshot stack */