Ramblings » Blog Archive » a bug on failure failure

I blogged about MySQL Bug 17928 a little while ago. Well, I’ve submitted a patch that fixes the problem. I wrote rather a rather detailed explanation in the Changeset comments (and I encourage every body who commits code anywhere to do the same). You can see the patch over on the commits list (or here if you don’t want to sift through the archives).

In all theory it shouldn’t be hard to enable multiple simultaneous backups for Cluster. How useful this would be is very debatable. Arguably of little use of all (it’s a REDO log backup). Of course, testing for this (as we test Node Failure) would start to get horrific. Any good arguments one way or the other are welcome.