siddontang

please send me a PR to fix it. Thanks:-)

多谢，应该是这句导致的 ``` tv[i] = append(tv[i], t) ``` 那如果是这样，我觉得可以在append的时候判断是不是capability不够了需要扩容，如果扩容了，可以干的事情是先重新分配一个两倍的slice，然后将原先的东西copy过去，在重新append，这样应该数据就是正确地了。 不过我现在还没测试，不知道是否可行。

Thank you @wllenyj Do you think what is the best way to fix it? Maybe Lock?

can you send me a PR to fix it?

can you send me a PR to fix it?

Hi @ongardie The LogCabin passes Jepsen test, but I don't test membership change. I want to test with LogCabin protocol instead of CLI client later, which means I should first...

maybe we can contribute it to etcd? /cc @xiang90 @tbg

@NingLin-P have we done any benchmark?

@BusyJay found a way to reduce the clone of entries, but the benchmark result was not improved largely. Please send the PR @BusyJay

We can name this feature Follower snapshot :-) Maybe we can even support Follower replication 🤔