Comment 6 for bug 1134892

@Shadow,

Thanks for the details.

Since you are able to reproduce this easily, can you provide
either/or both of the following:

1) A backtrace from the core file. You can use procedure here -- http://www.mysqlperformanceblog.com/2011/08/26/getting-mysql-core-file-on-linux/ to produce one. Make sure to add 'core-file-size=unlimited' under mysqld_safe in my.cnf as well (not mentioned there).

To produce backtrace from the core:

  gdb `which mysqld` --core ${core} --batch --quiet -ex "thread apply all bt full" -ex "quit"

  and attach that file here.

2) If you can provide a table(s), the statement and a minimal configuration on which it crashes, that would be for the best, since then I will be able to replicate it on my box and debug it better.

Since it doesn't crash without foreign keys, it certainly looks
like an issue with wsrep and not with PS, however, we need to
test so to reliably concur that.

@Shadow, are you able to repeat it with MySQL-wsrep (ie.
mysql+wsrep distributed by codership)?