diff --git a/src/test/cluster_tap/t/004_backend_types.pl b/src/test/cluster_tap/t/004_backend_types.pl
index 5c1c4b5366..376f8f302d 100644
--- a/src/test/cluster_tap/t/004_backend_types.pl
+++ b/src/test/cluster_tap/t/004_backend_types.pl
@@ -95,12 +95,21 @@
 	q{SELECT count(*) FROM pg_stat_activity
 	   WHERE backend_type IN (
 	       'heartbeat', 'interconnect listener',
-	       'lms worker',
 	       'managed recovery process', 'recovery coordinator',
 	       'recovery worker', 'sinval broadcaster', 'tt gc')});
 is($pgrac_visible, '0',
 	'no pgrac process descriptor visible except spawned skeletons (others deferred to Stage 2-6)');
 
+# spec-7.3 D2: the LMS DATA-plane worker pool ships with a default of
+# cluster.lms_workers = 2 -- worker 0 stays the historic 'lms' process and
+# workers 1..N-1 run as 'lms worker' aux processes, so exactly ONE
+# 'lms worker' descriptor is a spawned skeleton now (not deferred).
+my $lms_worker_visible = $node->safe_psql(
+	'postgres',
+	q{SELECT count(*) FROM pg_stat_activity WHERE backend_type = 'lms worker'});
+is($lms_worker_visible, '1',
+	'default LMS pool spawns exactly one lms worker aux process (spec-7.3 D2)');
+
 # LMON is spawned by postmaster (spec-1.11 Sprint A).  Verify it
 # appears in pg_stat_activity exactly once.
 my $lmon_visible = $node->safe_psql(
diff --git a/src/test/cluster_tap/t/078_ic_tier1_partial_send_recv.pl b/src/test/cluster_tap/t/078_ic_tier1_partial_send_recv.pl
index 823cbd5e76..cdf69fab2f 100644
--- a/src/test/cluster_tap/t/078_ic_tier1_partial_send_recv.pl
+++ b/src/test/cluster_tap/t/078_ic_tier1_partial_send_recv.pl
@@ -61,6 +61,12 @@ sub build_hello
 $node->append_conf('postgresql.conf', "cluster.enabled = on\n");
 $node->append_conf('postgresql.conf', "cluster.interconnect_tier = tier1\n");
 $node->append_conf('postgresql.conf', "cluster.node_id = 0\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$node->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 $node->append_conf('postgresql.conf', "cluster.allow_single_node = on\n");
 
 # Declare 2 nodes so we can spoof "peer 1" connecting to "peer 0" (us).
diff --git a/src/test/cluster_tap/t/288_cf_enqueue_concurrent.pl b/src/test/cluster_tap/t/288_cf_enqueue_concurrent.pl
index a4f14e62b3..f83f97ffe6 100644
--- a/src/test/cluster_tap/t/288_cf_enqueue_concurrent.pl
+++ b/src/test/cluster_tap/t/288_cf_enqueue_concurrent.pl
@@ -126,6 +126,13 @@ sub cf_counter
 $node0->append_conf('postgresql.conf', "cluster.node_id = 0\n");
 $node1->append_conf('postgresql.conf', $common_conf);
 $node1->append_conf('postgresql.conf', "cluster.node_id = 1\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$node0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$node1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 
 my $pgrac_conf = <<EOC;
 [cluster]
diff --git a/src/test/cluster_tap/t/289_cf_bootstrap.pl b/src/test/cluster_tap/t/289_cf_bootstrap.pl
index 632c0c93ab..41e9476d88 100644
--- a/src/test/cluster_tap/t/289_cf_bootstrap.pl
+++ b/src/test/cluster_tap/t/289_cf_bootstrap.pl
@@ -163,6 +163,13 @@
 $node0->append_conf('postgresql.conf', "cluster.node_id = 0\n");
 $node1->append_conf('postgresql.conf', $common_conf);
 $node1->append_conf('postgresql.conf', "cluster.node_id = 1\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$node0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$node1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 
 my $pgrac_conf = <<EOC;
 [cluster]
diff --git a/src/test/cluster_tap/t/334_ic_rdma_soft_roce.pl b/src/test/cluster_tap/t/334_ic_rdma_soft_roce.pl
index 504ac321d5..753d5446de 100644
--- a/src/test/cluster_tap/t/334_ic_rdma_soft_roce.pl
+++ b/src/test/cluster_tap/t/334_ic_rdma_soft_roce.pl
@@ -89,6 +89,9 @@ sub command_exists
 for my $node ($pair->node0, $pair->node1)
 {
 	$node->adjust_conf('postgresql.conf', 'cluster.interconnect_tier', 'tier3');
+	# spec-7.3 merge: this rig rewrites pgrac.conf with ONE data port per
+	# node; pin the pool to one worker (N=1 = spec-7.2 topology identity).
+	$node->adjust_conf('postgresql.conf', 'cluster.lms_workers', '1');
 }
 
 sub write_rdma_pgrac_conf
diff --git a/src/test/cluster_tap/t/335_adg_two_thread_rfs_apply.pl b/src/test/cluster_tap/t/335_adg_two_thread_rfs_apply.pl
index 2e1c4a0a89..644dde0e83 100644
--- a/src/test/cluster_tap/t/335_adg_two_thread_rfs_apply.pl
+++ b/src/test/cluster_tap/t/335_adg_two_thread_rfs_apply.pl
@@ -291,6 +291,15 @@ sub wait_for_log_match_from
 	'cluster.adg_rfs_conninfos = ' . quote_conf($rfs_conninfos) . "\n");
 $standby1->append_conf('postgresql.conf', $standby_common_conf);
 $standby1->append_conf('postgresql.conf', "cluster.node_id = 1\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$primary0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$primary1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$standby0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$standby1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 write_pair_conf($standby0, $standby1, 'adg2_standby');
 
 $standby0->start;
diff --git a/src/test/cluster_tap/t/336_cluster_backup_pitr.pl b/src/test/cluster_tap/t/336_cluster_backup_pitr.pl
index 660ddb5f0f..26ed2fb77c 100644
--- a/src/test/cluster_tap/t/336_cluster_backup_pitr.pl
+++ b/src/test/cluster_tap/t/336_cluster_backup_pitr.pl
@@ -34,6 +34,7 @@
 $node->append_conf('postgresql.conf',
 	  "cluster.enabled = on\n"
 	. "cluster.node_id = 0\n"
+	. "cluster.lms_workers = 1\n"
 	. "cluster.allow_single_node = on\n"
 	. "wal_level = replica\n"
 	. "archive_mode = on\n"
@@ -466,6 +467,7 @@ sub configure_pair_restore
 	. "restore_command = 'cp $pair_archive/%f %p'\n"
 	. "cluster.enabled = on\n"
 	. "cluster.node_id = 0\n"
+	. "cluster.lms_workers = 1\n"
 	. "cluster.allow_single_node = on\n"
 	. "cluster.voting_disks = ''\n"
 	. "cluster.wal_threads_dir = ''\n"
@@ -605,6 +607,7 @@ sub wait_for_pair_restore_promotion
 $bad_target_node->append_conf('postgresql.conf',
 	  "cluster.enabled = on\n"
 	. "cluster.node_id = 0\n"
+	. "cluster.lms_workers = 1\n"
 	. "cluster.allow_single_node = on\n"
 	. "wal_level = replica\n"
 	. "cluster.recovery_target_scn = '0'\n");
@@ -624,6 +627,7 @@ sub wait_for_pair_restore_promotion
 $multi_target_node->append_conf('postgresql.conf',
 	  "cluster.enabled = on\n"
 	. "cluster.node_id = 0\n"
+	. "cluster.lms_workers = 1\n"
 	. "cluster.allow_single_node = on\n"
 	. "wal_level = replica\n"
 	. "cluster.recovery_target_scn = '1'\n"
@@ -645,6 +649,7 @@ sub wait_for_pair_restore_promotion
 $pin_node->append_conf('postgresql.conf',
 	  "cluster.enabled = on\n"
 	. "cluster.node_id = 0\n"
+	. "cluster.lms_workers = 1\n"
 	. "cluster.allow_single_node = on\n"
 	. "wal_level = replica\n"
 	. "archive_mode = on\n"
diff --git a/src/test/cluster_tap/t/337_shared_catalog_ddl_2node.pl b/src/test/cluster_tap/t/337_shared_catalog_ddl_2node.pl
index 6db3cccd8b..26c4881a64 100644
--- a/src/test/cluster_tap/t/337_shared_catalog_ddl_2node.pl
+++ b/src/test/cluster_tap/t/337_shared_catalog_ddl_2node.pl
@@ -201,6 +201,13 @@
 $node1->append_conf('postgresql.conf', $sc_common);
 $node1->append_conf('postgresql.conf', $cluster_conf);
 $node1->append_conf('postgresql.conf', "cluster.node_id = 1\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$node0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$node1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 
 my $pgrac_conf = <<EOC;
 [cluster]
diff --git a/src/test/cluster_tap/t/339_shared_catalog_faults_2node.pl b/src/test/cluster_tap/t/339_shared_catalog_faults_2node.pl
index e6571dacd2..307fb57154 100644
--- a/src/test/cluster_tap/t/339_shared_catalog_faults_2node.pl
+++ b/src/test/cluster_tap/t/339_shared_catalog_faults_2node.pl
@@ -221,6 +221,13 @@
 $node1->append_conf('postgresql.conf', $sc_common);
 $node1->append_conf('postgresql.conf', $cluster_conf);
 $node1->append_conf('postgresql.conf', "cluster.node_id = 1\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$node0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$node1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 
 my $pgrac_conf = <<EOC;
 [cluster]
diff --git a/src/test/cluster_tap/t/346_shared_catalog_relmap_crash_2node.pl b/src/test/cluster_tap/t/346_shared_catalog_relmap_crash_2node.pl
index cd340fa646..bc8a7e64f2 100644
--- a/src/test/cluster_tap/t/346_shared_catalog_relmap_crash_2node.pl
+++ b/src/test/cluster_tap/t/346_shared_catalog_relmap_crash_2node.pl
@@ -185,6 +185,13 @@
 $node1->append_conf('postgresql.conf', $sc_common);
 $node1->append_conf('postgresql.conf', $cluster_conf);
 $node1->append_conf('postgresql.conf', "cluster.node_id = 1\n");
+# spec-7.3 merge: this hand-rolled rig reserves ONE data port per node; the
+# shipped default cluster.lms_workers=2 binds [data_port, data_port+1] and
+# cross-wires consecutive free ports (HELLO DATA worker mismatch).  Pin the
+# pool to one worker: N=1 is the spec-7.2 topology identity this rig was
+# written against.
+$node0->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
+$node1->append_conf('postgresql.conf', "cluster.lms_workers = 1\n");
 
 my $pgrac_conf = <<EOC;
 [cluster]