From 519d20cfe9cabe1611739e1a0ceff42ad62b261d Mon Sep 17 00:00:00 2001
From: Zijie <foreyes_1002@163.com>
Date: Wed, 10 Apr 2024 18:02:43 +0800
Subject: [PATCH] Add GUC 'gp_random_insert_segments' to control the segments
 used for random distributed table insertion

Introduces the 'gp_random_insert_segments' GUC to reduce the generation of
excessive fragmented files during the insertion of small amounts of data into
clusters with a large number of segments (e.g., 1000 records into 100 segments).

Fragmented data insertion can significantly degrade performance, especially
when using append-optimized or cloud-based storage. By introducing
the 'gp_random_insert_segments' GUC, users can limit the number of segments
used for data insertion in randomly distributed tables, which can significantly
reduce fragmented files.
---
 src/backend/cdb/cdbpath.c         |  9 +++++++++
 src/backend/commands/copyfrom.c   |  6 ++++++
 src/backend/utils/misc/guc_gp.c   | 11 +++++++++++
 src/include/utils/guc.h           |  1 +
 src/include/utils/sync_guc_name.h |  1 +
 5 files changed, 28 insertions(+)

diff --git a/src/backend/cdb/cdbpath.c b/src/backend/cdb/cdbpath.c
index 4fb64912cc0..7c8547e357d 100644
--- a/src/backend/cdb/cdbpath.c
+++ b/src/backend/cdb/cdbpath.c
@@ -2611,6 +2611,15 @@ create_motion_path_for_insert(PlannerInfo *root, GpPolicy *policy,
 	}
 	else
 		elog(ERROR, "unrecognized policy type %u", policyType);
+
+	if (CdbPathLocus_IsStrewn(subpath->locus) &&
+		gp_random_insert_segments > 0 &&
+		gp_random_insert_segments < CdbPathLocus_NumSegments(subpath->locus))
+	{
+		/* Select limited random segments for data insertion. */
+		subpath->locus.numsegments = gp_random_insert_segments;
+	}
+
 	return subpath;
 }
 
diff --git a/src/backend/commands/copyfrom.c b/src/backend/commands/copyfrom.c
index acb01d91ec6..9e02abf749f 100644
--- a/src/backend/commands/copyfrom.c
+++ b/src/backend/commands/copyfrom.c
@@ -3324,6 +3324,12 @@ GetTargetSeg(GpDistributionData *distData, TupleTableSlot *slot)
 
 		target_seg = cdbhashreduce(cdbHash); /* hash result segment */
 	}
+	else if (gp_random_insert_segments > 0 &&
+			 gp_random_insert_segments < policy->numsegments)
+	{
+		/* Select limited random segments for data insertion. */
+		target_seg = cdbhashrandomseg(gp_random_insert_segments);
+	}
 	else
 	{
 		/*
diff --git a/src/backend/utils/misc/guc_gp.c b/src/backend/utils/misc/guc_gp.c
index f7a1dd2b5cb..9a4a85feb5e 100644
--- a/src/backend/utils/misc/guc_gp.c
+++ b/src/backend/utils/misc/guc_gp.c
@@ -146,6 +146,7 @@ int			gp_appendonly_compaction_threshold = 0;
 bool		enable_parallel = false;
 int			gp_appendonly_insert_files = 0;
 int			gp_appendonly_insert_files_tuples_range = 0;
+int			gp_random_insert_segments = 0;
 bool		gp_heap_require_relhasoids_match = true;
 bool		gp_local_distributed_cache_stats = false;
 bool		debug_xlog_record_read = false;
@@ -3215,6 +3216,16 @@ struct config_int ConfigureNamesInt_gp[] =
 		NULL, NULL, NULL
 	},
 
+	{
+		{"gp_random_insert_segments", PGC_USERSET, CUSTOM_OPTIONS,
+			gettext_noop("Use limited number of segments for random distributed table insertion."),
+			NULL
+		},
+		&gp_random_insert_segments,
+		0, 0, INT_MAX,
+		NULL, NULL, NULL
+	},
+
 	{
 		{"gp_workfile_max_entries", PGC_POSTMASTER, RESOURCES,
 			gettext_noop("Sets the maximum number of entries that can be stored in the workfile directory"),
diff --git a/src/include/utils/guc.h b/src/include/utils/guc.h
index 90f054e7f33..d49f00a5bab 100644
--- a/src/include/utils/guc.h
+++ b/src/include/utils/guc.h
@@ -300,6 +300,7 @@ extern bool gp_appendonly_compaction;
 extern bool enable_parallel;
 extern int  gp_appendonly_insert_files;
 extern int  gp_appendonly_insert_files_tuples_range;
+extern int  gp_random_insert_segments;
 extern bool enable_answer_query_using_materialized_views;
 extern bool enable_offload_entry_to_qe;
 /*
diff --git a/src/include/utils/sync_guc_name.h b/src/include/utils/sync_guc_name.h
index aefc91e8716..14c5125a2d2 100644
--- a/src/include/utils/sync_guc_name.h
+++ b/src/include/utils/sync_guc_name.h
@@ -152,3 +152,4 @@
 		"gp_resgroup_debug_wait_queue",
 		"gp_appendonly_insert_files",
 		"gp_appendonly_insert_files_tuples_range",
+		"gp_random_insert_segments",