On 07/26/2012 04:06 PM, Dietmar Maurer wrote: > But recovery and cleanup actions can take several hours, so it is quite hard to find a window > on such system? We are always optimizing the recovery performance. For now, 30 nodes with dozens of TB data, the recovery process is less than 30 mins. Note, recovery can be nested, this means subsequent node event will supersede the previous one. This means, if you have 2 nodes failed one after another, the total time is: t0 + t(r), t0 is the window between these two event, and t(r) is the one node event recovery time. So yea, theoretically we can't assure mathematically recovery time is bound into a short window, but when it is reported recovery takes hours, I think it is time for us to revisit the code and make it faster. Thanks, Yuan |