by tlink » Fri Mar 26, 2010 1:23 pm
Yes, in this particular case both nodes did fail. I'm not sure why both failed. I purposely made one fail to test what would happen, but for some reason both of them "went down". Anyway I implemented your fix and it did work. Now here's another question. Wouldn't it be a really nice feature for the software to just do that for me? Not that it would should go ahead and do the resynch (I assume there's some risk of data corruption when this happens), but couldn't it easily analyze the two images for me and reccomend the best solution and give the option to implement?