前几天,Percona 资深工程师 Marco Tusa 爆料,升级到 MySQL 8.0.38 版本后,当实例中的表个数超过一万个,实例重启后会发生 Crash 而失败,即便是重启时加上 validate_tablespace_paths=OFF
也不行。
这个问题在 >= 8.0.38 版本中存在,包括 8.4.1 和 9.0.0。
详细复现过程参见:https://perconadev.atlassian.net/browse/PS-9306。
我用 MySQL 9.0.0 版本测试:
-- 创建一个最简单的表,并写入数据
> CREATE DATABASE test;
> USE test;
> CREATE TABLE t_1 (
`id` int NOT NULL,
PRIMARY KEY (`id`)
);
> INSERT INTO t_1 SELECT 1;
然后反复创建类似上面的表,表个数达到 1 万。
> SELECT COUNT(*) FROM information_schema.tables WHERE TABLE_SCHEMA ='test';
+----------+
| count(*) |
+----------+
| 10000 |
+----------+
之后重启实例,就能看到日志里有类似下面的内容,启动失败:
[Note] [MY-012207] [InnoDB] Using 2 threads to scan 10002 tablespace files
[Note] [MY-012200] [InnoDB] Thread# 0 - Checked 876/10002 files
...
[Note] [MY-012201] [InnoDB] Checked 10002 files
[Note] [MY-012208] [InnoDB] Completed space ID check of 10004 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.
BuildID[sha1]=7f06a4743d7801096bd81bc999201fdbca43a12c
Thread pointer: 0x0
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0 thread_stack 0x100000
[root@db160 mysql-9.0.0-linux-glibc2.17-x86_64-minimal]# #0 0x103f726 <unknown>
#1 0x103fa8c <unknown>
#2 0x7f18f666ac1f <unknown> at sysdeps/unix/sysv/linux/x86_64/sigaction.c:0
#3 0x218a7be <unknown>
#4 0x21705a7 <unknown>
#5 0x2b1d263 <unknown>
#6 0x7f18f6660179 start_thread at /usr/src/debug/glibc-2.28/nptl/pthread_create.c:479
#7 0x7f18f4811dc2 <unknown> at sysdeps/unix/sysv/linux/x86_64/clone.S:95
#8 0xffffffffffffffff <unknown>
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.
确实挺拉胯的。
另外,还是有办法可以规避的,也就是采用 共享/通用 表空间方案,例如:
-- 1. 共享表空间方案
> SET GLOBAL innodb_file_per_table = 0;
> CREATE TABLE ...;
-- 2. 通用表空间方案
> CREATE TABLESPACE test ADD DATAFILE 'test.ibd';
> CREATE TABLE t_1(...) TABLESPACE=test;
这个 Bug 很 low,但这个问题很小,也很容易规避
评论区