Why MySQL 8.0.38 Crashes with Over 10,000 Tables and How to Fix It

Percona senior engineer Marco Tusa reported that MySQL versions ≥ 8.0.38 (including 8.4.1 and 9.0.0) crash on instance restart when the number of tables exceeds ten thousand, and the usual workaround validate_tablespace_paths=OFF does not help.

Reproduction steps (tested on MySQL 9.0.0):

-- Create a simple database and table, then insert a row
CREATE DATABASE test;
USE test;
CREATE TABLE t_1 (
  `id` int NOT NULL,
  PRIMARY KEY (`id`)
);
INSERT INTO t_1 SELECT 1;

Repeatedly create similar tables until the count reaches 10,000:

SELECT COUNT(*) FROM information_schema.tables WHERE TABLE_SCHEMA='test';
+----------+
| count(*) |
+----------+
|    10000 |
+----------+

After restarting the instance, the error log shows InnoDB scanning thousands of tablespace files and eventually a segmentation fault (SIGSEGV), e.g.:

[Note] [MY-012207] [InnoDB] Using 2 threads to scan 10002 tablespace files
[Note] [MY-012200] [InnoDB] Thread # 0 - Checked 876/10002 files
...
2024-07-12T06:48:14Z UTC - mysqld got signal 11 ;
Signal SIGSEGV (Address not mapped to object) at address 0x508
Most likely, you have hit a bug, but this error can also be caused by malfunctioning hardware.
...

The bug is considered low‑severity but reproducible and easy to avoid. The recommended mitigation is to use a shared or general tablespace instead of the default per‑table tablespace:

-- 1. Shared tablespace (disable per‑table file)
SET GLOBAL innodb_file_per_table = 0;
CREATE TABLE ...;

-- 2. General tablespace
CREATE TABLESPACE test ADD DATAFILE 'test.ibd';
CREATE TABLE t_1 (...) TABLESPACE=test;

Both approaches have been verified to prevent the crash. Although the bug is minor, applying a shared/general tablespace configuration is a simple and effective workaround.