【brpc学习实践六】backup request场景案例

应用场景

有时为了保证可用性，需要同时访问两路服务，哪个先返回就取哪个。在brpc中，这有多种做法，根据server是否挂在同一个命名服务内有所区别。

当后端server可以挂在一个命名服务内时

Channel开启backup request。这个Channel会先向其中一个server发送请求，如果在ChannelOptions.backup_request_ms后还没回来，再向另一个server发送。之后哪个先回来就取哪个。在设置了合理的backup_request_ms后，大部分时候只会发一个请求，对后端服务只有一倍压力。

#include <gflags/gflags.h>
#include <butil/logging.h>
#include <butil/time.h>
#include <brpc/channel.h>
#include "echo.pb.h"DEFINE_string(protocol, "baidu_std", "Protocol type. Defined in src/brpc/options.proto");
DEFINE_string(connection_type, "", "Connection type. Available values: single, pooled, short");
DEFINE_string(server, "0.0.0.0:8000", "IP Address of server");
DEFINE_string(load_balancer, "", "The algorithm for load balancing");
DEFINE_int32(timeout_ms, 100, "RPC timeout in milliseconds");
DEFINE_int32(max_retry, 3, "Max retries(not including the first RPC)");
DEFINE_int32(backup_request_ms, 2, "Timeout for sending backup request");int main(int argc, char* argv[]) {// Parse gflags. We recommend you to use gflags as well.GFLAGS_NS::ParseCommandLineFlags(&argc, &argv, true);// A Channel represents a communication line to a Server. Notice that // Channel is thread-safe and can be shared by all threads in your program.brpc::Channel channel;// Initialize the channel, NULL means using default options.brpc::ChannelOptions options;options.protocol = FLAGS_protocol;options.connection_type = FLAGS_connection_type;options.timeout_ms = FLAGS_timeout_ms/*milliseconds*/;options.max_retry = FLAGS_max_retry;// 设置backup_request_ms参数，在超时后会自动向另外一个server重发请求，与首次请求一起，哪个先返回用哪个options.backup_request_ms = FLAGS_backup_request_ms;if (channel.Init(FLAGS_server.c_str(), FLAGS_load_balancer.c_str(), &options) != 0) {LOG(ERROR) << "Fail to initialize channel";return -1;}// Normally, you should not call a Channel directly, but instead construct// a stub Service wrapping it. stub can be shared by all threads as well.example::EchoService_Stub stub(&channel);// Send a request and wait for the response every 1 second.int counter = 0;while (!brpc::IsAskedToQuit()) {// We will receive response synchronously, safe to put variables// on stack.example::EchoRequest request;example::EchoResponse response;brpc::Controller cntl;request.set_index(++counter);// Because `done'(last parameter) is NULL, this function waits until// the response comes back or error occurs(including timedout).stub.Echo(&cntl, &request, &response, NULL);if (!cntl.Failed()) {LOG(INFO) << "Received response[index=" << response.index()<< "] from " << cntl.remote_side()<< " to " << cntl.local_side()<< " latency=" << cntl.latency_us() << "us";} else {LOG(WARNING) << cntl.ErrorText();}sleep(1);}LOG(INFO) << "EchoClient is going to quit";return 0;
}

backup_request_ms推荐值

backup_request_ms=2ms可以大约覆盖95.5%的请求，选择backup_request_ms=10ms则可以覆盖99.99%的请求。

当后端server不能挂在一个命名服务内时

方案一、使用SelectiveChannel：

建立一个开启backup request的SelectiveChannel，其中包含两个sub channel。访问这个SelectiveChannel和上面的情况类似，会先访问一个sub channel，如果在ChannelOptions.backup_request_ms后没返回，再访问另一个sub channel。如果一个sub channel对应一个集群，这个方法就是在两个集群间做互备。

#include <gflags/gflags.h>
#include <bthread/bthread.h>
#include <butil/logging.h>
#include <brpc/selective_channel.h>
#include <brpc/parallel_channel.h>
#include "echo.pb.h"DEFINE_int32(thread_num, 50, "Number of threads to send requests");
DEFINE_bool(use_bthread, false, "Use bthread to send requests");
DEFINE_int32(attachment_size, 0, "Carry so many byte attachment along with requests");
DEFINE_int32(request_size, 16, "Bytes of each request");
DEFINE_string(connection_type, "", "Connection type. Available values: single, pooled, short");
DEFINE_string(protocol, "baidu_std", "Protocol type. Defined in src/brpc/options.proto");
DEFINE_string(starting_server, "0.0.0.0:8114", "IP Address of the first server, port of i-th server is `first-port + i'");
DEFINE_string(load_balancer, "rr", "Name of load balancer");
DEFINE_int32