At Tue, 05 Jun 2012 13:57:22 +0800, Liu Yuan wrote: > > On 06/05/2012 11:40 AM, MORITA Kazutaka wrote: > > > Currently, SD_OP_GET_OBJ_LIST requests whole object ids though the > > most of them are discarded in screen_obj_list(). Can we optimize it > > and reduce the overhead dramatically? It is quite a big change to > > modify the recovery algorithm from pull-style to push-style. If > > possible, I'd like to avoid it. > > > I don't think the bulk of overhead is from screen_object_list(). Anyway, > I am kind of busy with fixing other bugs. I think we could get good > enough info which is biggest culprit by tracing. For last time I tried > with trace, fill_obj_list() took most of the time, which tried to > connect_to() all the nodes. If connect is the bottleneck, caching fds looks a simpler approach. Thanks, Kazutaka |