MySQL Bugs: #113100: [FATAL] fsync() returned EIO,aborting.[InnoDB] Assertion failure: ut0ut.cc:532

Bug #113100	[FATAL] fsync() returned EIO,aborting.[InnoDB] Assertion failure: ut0ut.cc:532
Submitted:	16 Nov 2023 8:52	Modified:	16 Nov 2023 13:06
Reporter:	oraunix xiang	Email Updates:
Status:	Can't repeat	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	8.0.18	OS:	Red Hat (7.6)
Assigned to:		CPU Architecture:	x86 (x64)
Tags:	UTC - mysqld got signal 6 ;

Description:
2023-11-14T10:19:01.070521+08:00 0[MY-0118251T [FATAL] fsync() returned EIO,aborting
2023-11-1410:19:01.070582+08:00 0 (ERROR) MY-013183] [InnoDB] Assertion failure: ut0ut.cc:532 thread 140224157959936

:02:19:01 UTC - mysqld got signal 6 ;
stack bottom = 0 thread stack 0x46000/mysql/bin/mysqld(my print stacktrace(unsigned char const*, unsigned long)+0x2e) [0xled6ccel
/usr/local/mysql/bin/mysqld(handle fatal signal+0x323) [0xfb2d23]

/mysgl/bin/mysqld(ib::fatal:;~fatal()+0x95) [0x2171ca5]
bin/mysqld(os_file_flush_func(int)+0x46b) [0x205a29b]
bin/mysqld(Fil_shard::redo_space_flush()+0x40d) [0x2296afdl
mysql/bin/mysqld(Fil_shard::flush_file_redo()+0x3a) [0x2296d2a]
mysgl/bin/mysgld(log_flusher(log_t*)+0x360) [0x204e440]
/usr/local/mysql/bin/mysgld/std::thread::_State_impl<std::thread::_Invoker<std::tuple<Runable, void (*)(log_t*), log_t*> > >::_M_run())+0xa9) [0x2023f19]

How to repeat:
This is a multi-instance environment, one of the instances reported an error, the other two instances are running normally, this problem instance has a crash message every once in a while, recently more and more frequent, a few hours appear once, the error is exactly the same

HI Mr. xiang,

Thank you for your bug report.

However, let us inform you that this is a forum for the bug reports with fully repeatable test cases.

That involves having all the tables, with their rows, uploaded to our site. See how it is done in the "Files" tab above.

Next, you are using a very, very old release of 8.0. Current release is 8.0.35.

Last, but not least, if the assert happens in the flushing operation, then this is most likely a hardware error or a glitch that happened in some RAM modules. This is usually the case when RAM modules are not of high quality, like those that have ECC RAM modules with redundant bits for RAM error checking and correcting.

Can't repeat.