agora inbox for [email protected]  

help / color / mirror / Atom feed

[PATCH v26a 1/2] instrumentation: Show additional TSC clock source info in pg_test_timing
346+ messages / 2 participants
[nested] [flat]

* [PATCH v26a 1/2] instrumentation: Show additional TSC clock source info in pg_test_timing
@ 2026-04-08 17:11  Lukas Fittl <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Lukas Fittl @ 2026-04-08 17:11 UTC (permalink / raw)

In some cases its needed to understand whether TSC frequency data was
sourced from CPUID, and which of the registers. This shows this debug
information at the end of pg_test_timing, through use of a new
pg_timing_tsc_clock_source_info function, replacing the previous
pg_tsc_calibrate_frequency export that was only needed for debug info.

Author: Lukas Fittl <[email protected]>
Suggested-by: Andres Freund <[email protected]>
Discussion: https://www.postgresql.org/message-id/flat/jr4hk2sxhqcfpb67ftz5g4vw33nm67cgf7go3wwmqsafu5aclq%405m67...
---
 src/include/port/pg_cpu.h               |  2 +-
 src/include/portability/instr_time.h    | 11 ++++++-
 src/common/instr_time.c                 | 44 +++++++++++++++++++++++--
 src/port/pg_cpu_x86.c                   | 23 +++++++++++--
 src/bin/pg_test_timing/pg_test_timing.c | 12 ++++---
 src/tools/pgindent/typedefs.list        |  1 +
 6 files changed, 81 insertions(+), 12 deletions(-)

diff --git a/src/include/port/pg_cpu.h b/src/include/port/pg_cpu.h
index a5d42f1b68d..db4cd62f7ef 100644
--- a/src/include/port/pg_cpu.h
+++ b/src/include/port/pg_cpu.h
@@ -56,7 +56,7 @@ x86_feature_available(X86FeatureId feature)
 	return X86Features[feature];
 }
 
-extern uint32 x86_tsc_frequency_khz(void);
+extern uint32 x86_tsc_frequency_khz(char *source, size_t source_len);
 
 #endif							/* defined(USE_SSE2) || defined(__i386__) */
 
diff --git a/src/include/portability/instr_time.h b/src/include/portability/instr_time.h
index 92558e234ac..4917728afc3 100644
--- a/src/include/portability/instr_time.h
+++ b/src/include/portability/instr_time.h
@@ -165,7 +165,16 @@ extern PGDLLIMPORT int32 timing_tsc_frequency_khz;
 
 extern void pg_initialize_timing_tsc(void);
 
-extern uint32 pg_tsc_calibrate_frequency(void);
+typedef struct TscClockSourceInfo
+{
+	int32		frequency_khz;	/* from CPUID or calibration */
+	int32		calibrated_frequency_khz;	/* from calibration loop, 0 if did
+											 * not converge, -1 if not yet run */
+	char		frequency_source[128];	/* describes how frequency was
+										 * determined */
+} TscClockSourceInfo;
+
+extern const TscClockSourceInfo *pg_timing_tsc_clock_source_info(void);
 
 #endif							/* PG_INSTR_TSC_CLOCK */
 
diff --git a/src/common/instr_time.c b/src/common/instr_time.c
index fc6e1852c30..2523d6a0df6 100644
--- a/src/common/instr_time.c
+++ b/src/common/instr_time.c
@@ -70,6 +70,8 @@ static void set_ticks_per_ns(void);
 static void set_ticks_per_ns_system(void);
 
 #if PG_INSTR_TSC_CLOCK
+static TscClockSourceInfo tsc_info = {.calibrated_frequency_khz = -1};
+
 static bool tsc_use_by_default(void);
 static void set_ticks_per_ns_for_tsc(void);
 #endif
@@ -166,6 +168,7 @@ set_ticks_per_ns_system(void)
 #if PG_INSTR_TSC_CLOCK
 
 static void tsc_detect_frequency(void);
+static uint32 pg_tsc_calibrate_frequency(void);
 
 /*
  * Initialize the TSC clock source by determining its usability and frequency.
@@ -202,21 +205,36 @@ static void
 tsc_detect_frequency(void)
 {
 	timing_tsc_frequency_khz = 0;
+	tsc_info.frequency_khz = 0;
+	tsc_info.frequency_source[0] = '\0';
 
 	/* We require RDTSCP support and an invariant TSC, bail if not available */
 	if (!x86_feature_available(PG_RDTSCP) || !x86_feature_available(PG_TSC_INVARIANT))
 		return;
 
 	/* Determine speed at which the TSC advances */
-	timing_tsc_frequency_khz = x86_tsc_frequency_khz();
+	timing_tsc_frequency_khz = x86_tsc_frequency_khz(tsc_info.frequency_source,
+													 sizeof(tsc_info.frequency_source));
 	if (timing_tsc_frequency_khz > 0)
+	{
+		tsc_info.frequency_khz = timing_tsc_frequency_khz;
 		return;
+	}
 
 	/*
 	 * CPUID did not give us the TSC frequency. We can instead measure the
 	 * frequency by comparing ticks against walltime in a calibration loop.
 	 */
-	timing_tsc_frequency_khz = pg_tsc_calibrate_frequency();
+	if (tsc_info.calibrated_frequency_khz < 0)
+		tsc_info.calibrated_frequency_khz = pg_tsc_calibrate_frequency();
+
+	timing_tsc_frequency_khz = tsc_info.calibrated_frequency_khz;
+	if (timing_tsc_frequency_khz > 0)
+	{
+		strlcpy(tsc_info.frequency_source, "x86, calibration",
+				sizeof(tsc_info.frequency_source));
+		tsc_info.frequency_khz = timing_tsc_frequency_khz;
+	}
 }
 
 /*
@@ -282,7 +300,7 @@ tsc_use_by_default(void)
 #define TSC_CALIBRATION_ITERATIONS	1000000
 #define TSC_CALIBRATION_SKIPS		100
 #define TSC_CALIBRATION_STABLE_CYCLES	10
-uint32
+static uint32
 pg_tsc_calibrate_frequency(void)
 {
 	instr_time	initial_wall;
@@ -369,4 +387,24 @@ pg_tsc_calibrate_frequency(void)
 	return (uint32) freq_khz;
 }
 
+/*
+ * Returns TSC clock source information for diagnostic purposes.
+ *
+ * On first call, runs the TSC calibration loop (if not already done during
+ * frequency detection) which may take up to TSC_CALIBRATION_MAX_NS.
+ * Subsequent calls return cached results.
+ *
+ * Note: This won't return the right info in EXEC_BACKEND builds if this were
+ * used in the backend (which it currently is not), as tsc_info is not copied
+ * using read_backend_variables - only the TSC frequency is.
+ */
+const TscClockSourceInfo *
+pg_timing_tsc_clock_source_info(void)
+{
+	if (tsc_info.calibrated_frequency_khz < 0)
+		tsc_info.calibrated_frequency_khz = pg_tsc_calibrate_frequency();
+
+	return &tsc_info;
+}
+
 #endif							/* PG_INSTR_TSC_CLOCK */
diff --git a/src/port/pg_cpu_x86.c b/src/port/pg_cpu_x86.c
index 150b4a1d574..c097807fc7f 100644
--- a/src/port/pg_cpu_x86.c
+++ b/src/port/pg_cpu_x86.c
@@ -13,7 +13,11 @@
  *-------------------------------------------------------------------------
  */
 
-#include "c.h"
+#ifndef FRONTEND
+#include "postgres.h"
+#else
+#include "postgres_fe.h"
+#endif
 
 #if defined(USE_SSE2) || defined(__i386__)
 
@@ -161,10 +165,13 @@ static uint32 x86_hypervisor_tsc_frequency_khz(void);
  * 0 indicates the frequency information was not accessible via CPUID.
  */
 uint32
-x86_tsc_frequency_khz(void)
+x86_tsc_frequency_khz(char *source, size_t source_len)
 {
 	unsigned int reg[4] = {0};
 
+	if (source)
+		strlcpy(source, "x86", source_len);
+
 	/*
 	 * If we're inside a virtual machine, try to fetch the TSC frequency from
 	 * the hypervisor, using a hypervisor specific method.
@@ -174,7 +181,11 @@ x86_tsc_frequency_khz(void)
 	 * to be wildly incorrect when virtualized.
 	 */
 	if (x86_feature_available(PG_HYPERVISOR))
+	{
+		if (source)
+			strlcat(source, ", hypervisor, cpuid 0x40000010", source_len);
 		return x86_hypervisor_tsc_frequency_khz();
+	}
 
 	/*
 	 * On modern Intel CPUs, the TSC is implemented by invariant timekeeping
@@ -207,6 +218,9 @@ x86_tsc_frequency_khz(void)
 		if (reg[EAX] == 0 || reg[EBX] == 0)
 			return 0;
 
+		if (source)
+			strlcat(source, ", cpuid 0x15", source_len);
+
 		return reg[ECX] / 1000 * reg[EBX] / reg[EAX];
 	}
 
@@ -217,7 +231,12 @@ x86_tsc_frequency_khz(void)
 	 */
 	pg_cpuid(0x16, reg);
 	if (reg[EAX] > 0)
+	{
+		if (source)
+			strlcat(source, ", cpuid 0x16", source_len);
+
 		return reg[EAX] * 1000;
+	}
 
 	return 0;
 }
diff --git a/src/bin/pg_test_timing/pg_test_timing.c b/src/bin/pg_test_timing/pg_test_timing.c
index 2afb0e6a410..ef5dd90c41f 100644
--- a/src/bin/pg_test_timing/pg_test_timing.c
+++ b/src/bin/pg_test_timing/pg_test_timing.c
@@ -184,7 +184,7 @@ static void
 test_tsc_timing(void)
 {
 	uint64		loop_count;
-	uint32		calibrated_freq;
+	const TscClockSourceInfo *info;
 
 	printf("\n");
 	loop_count = test_timing(test_duration, TIMING_CLOCK_SOURCE_TSC, false);
@@ -198,11 +198,13 @@ test_tsc_timing(void)
 		output(loop_count);
 		printf("\n");
 
-		printf(_("TSC frequency in use: %u kHz\n"), timing_tsc_frequency_khz);
+		info = pg_timing_tsc_clock_source_info();
+		printf(_("TSC frequency in use: %d kHz\n"), info->frequency_khz);
+		if (info->frequency_source[0] != '\0')
+			printf(_("TSC frequency source: %s\n"), info->frequency_source);
 
-		calibrated_freq = pg_tsc_calibrate_frequency();
-		if (calibrated_freq > 0)
-			printf(_("TSC frequency from calibration: %u kHz\n"), calibrated_freq);
+		if (info->calibrated_frequency_khz > 0)
+			printf(_("TSC frequency from calibration: %d kHz\n"), info->calibrated_frequency_khz);
 		else
 			printf(_("TSC calibration did not converge\n"));
 
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ea95e7984bc..2137f1e0cfe 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -3090,6 +3090,7 @@ TParserStateActionItem
 TQueueDestReceiver
 TRGM
 TSAnyCacheEntry
+TscClockSourceInfo
 TSConfigCacheEntry
 TSConfigInfo
 TSDictInfo
-- 
2.53.0.1.gb2826b52eb


--5ynaw2hkmtmijn2o
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
	filename="v26a-0002-fixup-instrumentation-Show-additional-TSC-clock.patch"



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread

* [PATCH 7/8] Decouple updating of freezing information from swap_relation_files().
@ 2026-06-16 11:54  Antonin Houska <[email protected]>
  0 siblings, 0 replies; 346+ messages in thread

From: Antonin Houska @ 2026-06-16 11:54 UTC (permalink / raw)

So far, the relfrozenxid and relminmxid attributes of pg_class used to be
updated by swap_relation_files(), however, not every call of the function
should do that. On the contrary, new feature of REPACK (CONCURRENTLY) will
need to update this information in one case that would be tricky to handle
using swap_relation_files().
---
 src/backend/commands/repack.c | 114 +++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 58 deletions(-)

diff --git a/src/backend/commands/repack.c b/src/backend/commands/repack.c
index 6f3e4ecff46..92c1b678e6f 100644
--- a/src/backend/commands/repack.c
+++ b/src/backend/commands/repack.c
@@ -186,6 +186,8 @@ static void copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldInde
 							TransactionId *pFreezeXid,
 							MultiXactId *pCutoffMulti,
 							ChangeContext *chgcxt);
+static void update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+									MultiXactId cutoffMulti);
 static List *get_tables_to_repack(RepackCommand cmd, bool usingindex,
 								  MemoryContext permcxt);
 static List *get_tables_to_repack_partitioned(RepackCommand cmd,
@@ -2014,13 +2016,6 @@ copy_table_data(Relation NewHeap, Relation OldHeap, Relation OldIndex,
  * links) while the latter is the only way to handle cases in which a toast
  * table is added or removed altogether.
  *
- * Additionally, the first relation is marked with relfrozenxid set to
- * frozenXid.  It seems a bit ugly to have this here, but the caller would
- * have to do it anyway, so having it here saves a heap_update.  Note: in
- * the swap-toast-links case, we assume we don't need to change the toast
- * table's relfrozenxid: the new version of the toast table should already
- * have relfrozenxid set to RecentXmin, which is good enough.
- *
  * Lastly, if r2 and its toast table and toast index (if any) are mapped,
  * their OIDs are emitted into mapped_tables[].  This is hacky but beats
  * having to look the information up again later in finish_heap_swap.
@@ -2029,9 +2024,7 @@ static void
 swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 					bool swap_toast_by_content,
 					bool is_internal,
-					TransactionId frozenXid,
-					MultiXactId cutoffMulti,
-					Oid *mapped_tables)
+					Oid *mapped_tables, Oid *r1_toastid)
 {
 	Relation	relRelation;
 	HeapTuple	reltup1,
@@ -2063,6 +2056,9 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	relam1 = relform1->relam;
 	relam2 = relform2->relam;
 
+	if (r1_toastid)
+		*r1_toastid = relform1->reltoastrelid;
+
 	if (RelFileNumberIsValid(relfilenumber1) &&
 		RelFileNumberIsValid(relfilenumber2))
 	{
@@ -2178,15 +2174,6 @@ 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 and minimum MultiXid */
-	if (relform1->relkind != RELKIND_INDEX)
-	{
-		Assert(!TransactionIdIsValid(frozenXid) ||
-			   TransactionIdIsNormal(frozenXid));
-		relform1->relfrozenxid = frozenXid;
-		relform1->relminmxid = cutoffMulti;
-	}
-
 	/* swap size statistics too, since new rel has freshly-updated stats */
 	{
 		int32		swap_pages;
@@ -2288,9 +2275,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 									target_is_pg_class,
 									swap_toast_by_content,
 									is_internal,
-									frozenXid,
-									cutoffMulti,
-									mapped_tables);
+									mapped_tables,
+									NULL);
 			}
 			else
 			{
@@ -2391,9 +2377,8 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 							target_is_pg_class,
 							swap_toast_by_content,
 							is_internal,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 	}
 
 	/* Clean up. */
@@ -2403,6 +2388,36 @@ swap_relation_files(Oid r1, Oid r2, bool target_is_pg_class,
 	table_close(relRelation, RowExclusiveLock);
 }
 
+/*
+ * Update relfrozenxid and relminmxid attributes of pg_class entry, specified
+ * by OID.
+ */
+static void
+update_relation_cutoffs(Oid relid, TransactionId frozenXid,
+						MultiXactId cutoffMulti)
+{
+	Relation	relRelation;
+	HeapTuple	reltup;
+	Form_pg_class relform;
+	CatalogIndexState indstate;
+
+	relRelation = table_open(RelationRelationId, RowExclusiveLock);
+	reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(relid));
+	if (!HeapTupleIsValid(reltup))
+		elog(ERROR, "cache lookup failed for relation %u", relid);
+	relform = (Form_pg_class) GETSTRUCT(reltup);
+	relform->relfrozenxid = frozenXid;
+	relform->relminmxid = cutoffMulti;
+
+	indstate = CatalogOpenIndexes(relRelation);
+	CatalogTupleUpdateWithInfo(relRelation, &reltup->t_self, reltup,
+							   indstate);
+	CatalogCloseIndexes(indstate);
+
+	heap_freetuple(reltup);
+	table_close(relRelation, RowExclusiveLock);
+}
+
 /*
  * Remove the transient table that was built by make_new_heap, and finish
  * cleaning up (including rebuilding all indexes on the old heap).
@@ -2420,6 +2435,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 {
 	ObjectAddress object;
 	Oid			mapped_tables[4];
+	Oid			oid_old_toastid;
 	int			i;
 
 	/* Report that we are now swapping relation files */
@@ -2436,7 +2452,7 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 	swap_relation_files(OIDOldHeap, OIDNewHeap,
 						(OIDOldHeap == RelationRelationId),
 						swap_toast_by_content, is_internal,
-						frozenXid, cutoffMulti, mapped_tables);
+						mapped_tables, &oid_old_toastid);
 
 	/*
 	 * If it's a system catalog, queue a sinval message to flush all catcaches
@@ -2492,37 +2508,20 @@ finish_heap_swap(Oid OIDOldHeap, Oid OIDNewHeap,
 								 PROGRESS_REPACK_PHASE_FINAL_CLEANUP);
 
 	/*
-	 * If the relation being rebuilt is pg_class, swap_relation_files()
-	 * couldn't update pg_class's own pg_class entry (check comments in
-	 * swap_relation_files()), thus relfrozenxid was not updated. That's
-	 * annoying because a potential reason for doing a VACUUM FULL is a
-	 * imminent or actual anti-wraparound shutdown.  So, now that we can
-	 * access the new relation using its indices, update relfrozenxid.
-	 * pg_class doesn't have a toast relation, so we don't need to update the
-	 * corresponding toast relation. Not that there's little point moving all
-	 * relfrozenxid updates here since swap_relation_files() needs to write to
-	 * pg_class for non-mapped relations anyway.
+	 * Update relfrozenxid and relminmxid. CCI is needed to see the most
+	 * recent pg_class tuple, created by swap_relation_files(). However that
+	 * CCI makes us use the new relation file. Therefore we cannot do this
+	 * before indexes have been rebuilt too.
 	 */
-	if (OIDOldHeap == RelationRelationId)
-	{
-		Relation	relRelation;
-		HeapTuple	reltup;
-		Form_pg_class relform;
-
-		relRelation = table_open(RelationRelationId, RowExclusiveLock);
-
-		reltup = SearchSysCacheCopy1(RELOID, ObjectIdGetDatum(OIDOldHeap));
-		if (!HeapTupleIsValid(reltup))
-			elog(ERROR, "cache lookup failed for relation %u", OIDOldHeap);
-		relform = (Form_pg_class) GETSTRUCT(reltup);
-
-		relform->relfrozenxid = frozenXid;
-		relform->relminmxid = cutoffMulti;
-
-		CatalogTupleUpdate(relRelation, &reltup->t_self, reltup);
+	CommandCounterIncrement();
+	update_relation_cutoffs(OIDOldHeap, frozenXid, cutoffMulti);
 
-		table_close(relRelation, RowExclusiveLock);
-	}
+	/*
+	 * The same for TOAST, if needed. In the swap-toast-links case, the new
+	 * the toast table should already have relfrozenxid set to RecentXmin.
+	 */
+	if (OidIsValid(oid_old_toastid) && swap_toast_by_content)
+		update_relation_cutoffs(oid_old_toastid, frozenXid, cutoffMulti);
 
 	/* Destroy new heap with old filenumber */
 	object.classId = RelationRelationId;
@@ -4072,9 +4071,8 @@ rebuild_relation_finish_concurrent(Relation NewHeap, Relation OldHeap,
 							(old_table_oid == RelationRelationId),
 							false,	/* swap_toast_by_content */
 							true,
-							InvalidTransactionId,
-							InvalidMultiXactId,
-							mapped_tables);
+							mapped_tables,
+							NULL);
 
 #ifdef USE_ASSERT_CHECKING
 
-- 
2.52.0


--=-=-=
Content-Type: text/plain
Content-Disposition: attachment;
 filename=v01-0008-Make-REPACK-CONCURRENTLY-MVCC-safe.patch



^ permalink  raw  reply  [nested|flat] 346+ messages in thread