Comment 6 for bug 1590166

Hi Paul
I've spent some time trying to reproduce it without success yet.

I was using galera cluster (three nodes) with scenario:
step one : put one node down
step two : create some flow of changes for data in db ( update/insert/delete )
step three: try to rejoin node.

If you will be able to provide me any additional information, it would be great. Maybe you can provide me some raw estimates about how many changes happened ( for example new instances launched) .

Thanks.