Re: index corruption on windows

[Ingo Strüwing]

Yes, I am aware of the cries. But no one by far could ever come up with something that would make an at least somewhat repeatable test case. Without this we could just speculate about the cause. Since it seems always have to do with heavy concurrency, one could point to our Windows implementation of thread locking. But unfortunately we are short of Windows system specialists. And even such a guy could easily miss a perhaps simple mistake without a repeatable test case.

So yes, I am very sorry. I need to "throw your message to the ring". I will however try to initiate an internal project for creating a heavy concurrency test system on Windows to hopefully get corruptions internally. But the outcome is uncertain.

One thing that you could try is to ALTER TABLE ... ENGINE=InnoDB or something for your tables. Other storage engines will not suffer from MyISAM corruptions. ;) But they might have other disadvantages though.