On Mon, Jul 30, 2012 at 3:06 PM, Liu Yuan <namei.unix at gmail.com> wrote: > On 07/29/2012 10:29 PM, Yunkai Zhang wrote: >> = Usage = >> 1) Start delay recovery transaction: >> $ collie delay_recovery start >> Delay recovery start ... >> >> 2) Add(kill) multiple nodes into(from) cluster >> ... >> >> 3) Stop delay recovery transaction: >> $ collie delay_recovery stop >> Delay recovery stop. > > I am not reading the code yet, I just have a question: can this process > survive a node failure event? For e.g, > > $ collie delay_recovery start > - some nodes go down while adding manually some nodes > ^ > | > how does you react to this scenario? Can you show more information to me? it works well in my testing, and I'm refactoring these series patchset, maybe we can test it later after I submit v2. > > Thanks, > Yuan -- Yunkai Zhang Work at Taobao |