1 files changed, 9 insertions, 1 deletions
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index d8345a75b68..c3b000c19b2 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -262,9 +262,17 @@ ReorderBufferAllocate(void)
 											SLAB_DEFAULT_BLOCK_SIZE,
 											sizeof(ReorderBufferTXN));
 
+	/*
+	 * To minimize memory fragmentation caused by long-running transactions
+	 * with changes spanning multiple memory blocks, we use a single
+	 * fixed-size memory block for decoded tuple storage. The performance
+	 * testing showed that the default memory block size maintains logical
+	 * decoding performance without causing fragmentation due to concurrent
+	 * transactions.
+	 */
 	buffer->tup_context = GenerationContextCreate(new_ctx,
 												  "Tuples",
-												  SLAB_LARGE_BLOCK_SIZE);
+												  SLAB_DEFAULT_BLOCK_SIZE);
 
 	hash_ctl.keysize = sizeof(TransactionId);
 	hash_ctl.entrysize = sizeof(ReorderBufferTXNByIdEnt);