From 675d6627103dde80ec6a2b048f102261149c6ced Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jean-Fran=C3=A7ois=20Gagn=C3=A9?=
 <8471576+jfg956@users.noreply.github.com>
Date: Tue, 3 Dec 2024 09:42:40 -0500
Subject: [PATCH] Contrib.

---
 storage/innobase/fil/fil0fil.cc       |  7 +++++--
 storage/innobase/handler/ha_innodb.cc | 24 ++++++++++++++++++++++++
 storage/innobase/include/srv0srv.h    |  4 ++++
 storage/innobase/srv/srv0srv.cc       |  4 ++++
 4 files changed, 37 insertions(+), 2 deletions(-)

diff --git a/storage/innobase/fil/fil0fil.cc b/storage/innobase/fil/fil0fil.cc
index 5b5bf18eca31..48d317c1d78f 100644
--- a/storage/innobase/fil/fil0fil.cc
+++ b/storage/innobase/fil/fil0fil.cc
@@ -11343,8 +11343,11 @@ dberr_t Tablespace_dirs::scan() {
   Space_id_set unique;
   Space_id_set duplicates;
 
-  /* Get the number of additional threads needed to scan the files. */
-  size_t n_threads = fil_get_scan_threads(ibd_files.size());
+  /* I removed a comment that was redundant, feel free to add it back if you think it added value.
+   * I left this comment for clarity of the patch, it can be removed when merging. */
+  size_t n_threads = (srv_tablespace_duplicate_check_threads == -1)
+                        ? fil_get_scan_threads(ibd_files.size())
+                        : srv_tablespace_duplicate_check_threads;
 
   if (n_threads > 0) {
     ib::info(ER_IB_MSG_382)
diff --git a/storage/innobase/handler/ha_innodb.cc b/storage/innobase/handler/ha_innodb.cc
index cf964ee2f4b7..b6aa8043455b 100644
--- a/storage/innobase/handler/ha_innodb.cc
+++ b/storage/innobase/handler/ha_innodb.cc
@@ -23439,6 +23439,29 @@ static MYSQL_SYSVAR_STR(directories, srv_innodb_directories,
                         "'innodb-data-home-dir;innodb-undo-directory;datadir'",
                         nullptr, nullptr, nullptr);
 
+/* there is no point in changing this during runtime, thus readonly */
+/* above copied from buffer_pool_load_at_startup */
+/* I / JFG thinks above predates SET PERSIST...
+ * it might make sense to change the value for the next startup. */
+/* TODO: explore allowing to change this with SET PERSIST (but not SET GLOBAL). */
+static MYSQL_SYSVAR_INT(
+    tablespace_duplicate_check_threads, srv_tablespace_duplicate_check_threads,
+    PLUGIN_VAR_OPCMDARG | PLUGIN_VAR_READONLY | PLUGIN_VAR_NOPERSIST,
+    "Number of threads for tablespace duplicate check at startup: "
+    "-1 for auto-compute, 0 for scaning in main thread, 1 in one sub-thread, "
+    "n in sub-threads with spilling in main thread",
+    nullptr, /* check */
+    nullptr, /* update */
+    -1, /* def */
+    -1, /* min */
+    512, /* max, there might be edge-cases for more than 512 threads, but enough for now.
+          * Note that we need 256 threads to leverage 256k iops with 1 milisecond IO latency,
+          *   that AWS io2 EBS volumes can have up to 256k iops with latencies between 0.3 and 0.5 ms,
+          *   so 512 threads is not such an unrealistic max value (maybe this should even be 2048). */
+    0 /* blk, unclear what this is, doc (link below) not helpful, copied from others */);
+/* doc link for blk above:
+ * https://dev.mysql.com/doc/extending-mysql/8.0/en/plugin-status-system-variables.html */
+
 #ifdef UNIV_DEBUG
 /** Use this variable innodb_interpreter to execute debug code within InnoDB.
 The output is stored in the innodb_interpreter_output variable. */
@@ -23597,6 +23620,7 @@ static SYS_VAR *innobase_system_variables[] = {
     MYSQL_SYSVAR(sort_buffer_size),
     MYSQL_SYSVAR(online_alter_log_max_size),
     MYSQL_SYSVAR(directories),
+    MYSQL_SYSVAR(tablespace_duplicate_check_threads),
     MYSQL_SYSVAR(sync_spin_loops),
     MYSQL_SYSVAR(spin_wait_delay),
     MYSQL_SYSVAR(spin_wait_pause_multiplier),
diff --git a/storage/innobase/include/srv0srv.h b/storage/innobase/include/srv0srv.h
index b6bf10810e00..28456310ee18 100644
--- a/storage/innobase/include/srv0srv.h
+++ b/storage/innobase/include/srv0srv.h
@@ -400,6 +400,10 @@ extern bool srv_numa_interleave;
 deliminated by ';', i.e the FIL_PATH_SEPARATOR. */
 extern char *srv_innodb_directories;
 
+/* Documented in storage/innobase/handler/ha_innodb.cc. */
+/* It looks pointless to duplicate comments, if needed, feel free to do it when merging. */
+extern int srv_tablespace_duplicate_check_threads;
+
 /** Server undo tablespaces directory, can be absolute path. */
 extern char *srv_undo_dir;
 
diff --git a/storage/innobase/srv/srv0srv.cc b/storage/innobase/srv/srv0srv.cc
index fb852a6fcd8d..acb86b7f0f83 100644
--- a/storage/innobase/srv/srv0srv.cc
+++ b/storage/innobase/srv/srv0srv.cc
@@ -146,6 +146,10 @@ char *srv_doublewrite_dir = nullptr;
 deliminated by ';', i.e the FIL_PATH_SEPARATOR. */
 char *srv_innodb_directories = nullptr;
 
+/* Documented in storage/innobase/handler/ha_innodb.cc. */
+/* It looks pointless to duplicate comments, if needed, feel free to fo it when merging. */
+int srv_tablespace_duplicate_check_threads;
+
 /** Number of threads spawned for initializing rollback segments
 in parallel */
 uint32_t srv_rseg_init_threads = 1;