简体繁体 English

gRPC：在C ++中关闭异步服务器的推荐方法是什么？

[英]gRPC: What is the recommended way to shut down an asynchronous server in C++?

原文 2016-02-29 19:50:11 1 2 c++/ asynchronous/ server/ grpc/ shutdown

I have a gRPC server that hosts two asynchronous services ("Master" and "Worker"), and I would like to implement graceful shutdown for the server. 我有一个gRPC服务器，它承载两个异步服务（“Master”和“Worker”），我想为服务器实现正常关闭。 Each service has its own grpc::CompletionQueue . 每个服务都有自己的grpc::CompletionQueue 。

There appear to be two Shutdown() methods that might be relevant: grpc::CompletionQueue::Shutdown() and grpc::Server::Shutdown() , but it's not clear from the documentation which ones should be used. 似乎有两个可能相关的Shutdown()方法： grpc::CompletionQueue::Shutdown()和grpc::Server::Shutdown() ，但是从文档中不清楚应该使用哪些方法。

What is a good pattern for shutting down an asynchronous service? 关闭异步服务有什么好的模式？

2 个解决方案

TL;DR: You must call both grpc::Server::Shutdown() and grpc::CompletionQueue::Shutdown() (for each completion queue used in the service) to shut down cleanly. TL; DR：你必须同时调用grpc::Server::Shutdown()和grpc::CompletionQueue::Shutdown() （对于服务中使用的每个完成队列）来干净地关闭。

If you call cq_->Shutdown() , the only observable effect is that subsequent calls to Service::AsyncService::RequestFoo() (the generated method for the corresponding Foo RPC) fail with an assertion. 如果你调用cq_->Shutdown() ，唯一可观察的效果是后续调用Service::AsyncService::RequestFoo() （相应的Foo RPC的生成方法）失败并带有断言。 From reading the documentation of the corresponding C API method ( grpc_completion_queue_shutdown() ), it appears that it is illegal to add new work to the queue—ie by calling RequestFoo() —so I added an is_shutdown_ member to my service wrapper classes (protected by a mutex) so that no enqueue attempts are made after cq_->Shutdown() is called. 通过阅读相应的C API方法（ grpc_completion_queue_shutdown() ）的文档，似乎将新工作添加到队列是非法的 - 即通过调用RequestFoo()我将is_shutdown_成员添加到我的服务包装类（protected）通过互斥锁），以便在调用cq_->Shutdown()之后不会进行入队尝试。 However, after doing this, the completion queue blocks indefinitely in cq_->Next() . 但是，执行此操作后，完成队列将无限期地阻塞在cq_->Next() 。 None of the enqueued tags complete (with an error or otherwise). 没有排队的标签完成（有错误或其他）。
If instead you call server_->Shutdown() , all of the enqueued tags complete immediately (with ok == false ). 如果您调用server_->Shutdown() ，则所有排队的标签立即完成（使用ok == false ）。 However, the completion queue continues to block indefinitely in cq_->Next() . 但是，完成队列继续在cq_->Next()无限期地阻塞。

Calling both cq_->Shutdown() (for each defined completion queue) and server_->Shutdown() results in a clean shutdown. 同时调用cq_->Shutdown() （对于每个已定义的完成队列）和server_->Shutdown()导致干净关闭。

One caveat: if you use grpc::ServerContext::AsyncNotifyWhenDone() to register a tag for call cancellation, these will not be returned by cq_->Next() if the server shuts down before the initial request is received for that call. 需要注意的是：如果使用grpc::ServerContext::AsyncNotifyWhenDone()来注册呼叫取消标签，那么如果服务器在收到该呼叫的初始请求之前关闭，则cq_->Next()将不会返回这些标签。 You will need to be cautious with the memory management of the corresponding tag structure, if you want to avoid memory leaks. 如果要避免内存泄漏，则需要对相应标记结构的内存管理保持谨慎。

Wait() void grpc::Server::Wait () Wait（）void grpc :: Server :: Wait（）
overridevirtual Block until the server shuts down. overridevirtual阻塞，直到服务器关闭。

Warning The server must be either shutting down or some other thread must call Shutdown for this function to ever return. 警告服务器必须要么关闭，要么其他线程必须调用Shutdown才能返回此功能。