Add cli.md client.md and example/README.md

Change-Id: Ibf9dc3cef49c923bd4ac62700a8b3871e082d4ba

docs/cn/cli.md

@@ -0,0 +1,36 @@
+braft提供了一系列API用来控制复制主或者具体节点, 可以选择在程序了调用[API](../../src/braft/cli.h)或者使用[braft_cli](../../tools/braft_cli.cpp)来给节点发远程控制命令
+
+#API
+
+```cpp
+// Add a new peer into the replicating group which consists of |conf|.
+// Returns OK on success, error infomation otherwise.
+butil::Status add_peer(const GroupId& group_id, const Configuration& conf,
+                       const PeerId& peer_id, const CliOptions& options);
+
+// Remove the peer from the replicating group which consists of |conf|.
+// Returns OK on success, error infomation otherwise.
+butil::Status remove_peer(const GroupId& group_id, const Configuration& conf,
+                          const PeerId& peer_id, const CliOptions& options);
+
+// Reset the configuration of the given peer
+butil::Status set_peer(const GroupId& group_id, const PeerId& peer_id,
+                       const Configuration& new_conf, const CliOptions& options);
+
+// Trigger snapshot of the peer
+butil::Status snapshot(const GroupId& group_id, const PeerId& peer_id,
+                       const CliOptions& options);
+```
+
+# braft_cli
+
+braft_cli提供了命令行工具, 作用和API类似
+
+```shell
+braft_cli: Usage: braft_cli [Command] [OPTIONS...]
+Command:
+  add_peer --group=$group_id --peer=$adding_peer --conf=$current_conf
+  remove_peer --group=$group_id --peer=$removing_peer --conf=$current_conf
+  set_peer --group=$group_id --peer==$target_peer --conf=$target_conf
+  snapshot --group=$group_id --peer=$target_peer
+```

docs/cn/client.md

@@ -0,0 +1,37 @@
+braft并不能直接被任何client访问， 本文主要是说明一个能访问braft节点的client需要那些要素。
+
+# Example
+
+[client side code](../../example/counter/client.cpp) of Counter
+
+# 总体流程
+
+要访问braft的主节点，需要做这么一些事情:
+
+* 需要知道这个复制组有哪些节点， 这个可以通过配置列表，记录在dns，或者提供某些naming service如集群的master，redis, zookeeper, etcd等。
+* 查询Leader位置
+* 感知Leader变化
+* 向Leader发起RPC.
+
+##RouteTable
+
+braft提供了[RouteTable](../../src/braft/route_table.h)功能，命名空间在braft::rtb, 可以帮助你的进程记录和追踪某个节点的主节点位置, 包含以下功能
+
+```cpp
+// Update configuration of group in route table
+int update_configuration(const GroupId& group, const Configuration& conf);
+int update_configuration(const GroupId& group, const std::string& conf_str);
+// Get the cached leader of group.
+// Returns:
+//  0 : success
+//  1 : Not sure about the leader
+//  -1, otherwise
+int select_leader(const GroupId& group, PeerId* leader);
+// Update leader
+int update_leader(const GroupId& group, const PeerId& leader);
+int update_leader(const GroupId& group, const std::string& leader_str);
+// Blocking the thread until query_leader finishes
+butil::Status refresh_leader(const GroupId& group, int timeout_ms);
+// Remove this group from route table
+int remove_group(const GroupId& group);
+```

example/README.md

@@ -0,0 +1,32 @@
+Brief introduction of examples:
+
+* Counter: A integer which can be added to by a given number in each request.
+* Atomic: : A integer thats supports exchange and compare_exchange operation.
+* Block: A block device supports random read/write from multiple threaded.
+
+# Build steps
+
+```shell
+example=counter|atomic|block
+cd $example && cmake . && make
+```
+
+# Run Server
+
+```sh
+sh run_server.sh
+```
+
+* Default number of servers in the group is `3`,  changed by `--server_num` 
+* Servers run on `./runtime/` which is resued,  add --clean to cleanup storage
+
+# Run Client
+
+```sh
+sh run_client.sh
+```
+
+* Default concurrency of client is 1, changed by `--thread_num` 
+* If server_num of run_server.sh has been changed, specify run_client to make it consistent.
+* Add `--log_each_request` if you want detailed information of each request.
+