Re: Replication crashes when max_bin_log_size is reached

[Guilhem Bichot]

Thanks, yes it is valid. But it says that crash happened in malloc() in the thread which creates threads to handle connections, which probably means that memory was badly corrupted before the crash (malloc() usually isn't crashing :). The crazy content of master.info was already a sign of that.
The error messages say that the code (SQL slave thread) which removes the old, useless relay log (just after the relay log rotation occured), fails to find the name of the next relay log.
A stupid question: are your relay logs placed on a NFS mount?

Mr. Guilhem Bichot <[e-mail]>
MySQL AB, Lead Software Engineer
Bordeaux, France
www.mysql.com