Interesting. What do you mean "it borrows"? gRPC protobuf compiler automatically generates the service routine/function for me. I just fill in my implementation/logic there. When a client request comes in this routine gets triggered. My code does not spawn any thread to call this service routine. It should be some gRPC low level IO receiver/read thread that calls into my service routine. Can you please check?
btw, can you please kindly answer my second question too? Carl Mastrangelo於 2017年1月4日星期三 UTC-8上午11時17分58秒寫道： > > From my understanding of the C-core and C++ wrapper for gRPC, it does not > create threads; it borrows. You must provide your own threads to handle > streaming messages. You can wait on multiple actions within a single > thread. > > On Tuesday, January 3, 2017 at 4:56:02 PM UTC-8, [email protected] > wrote: >> >> Thanks for your reply. >> >> I am using c++. >> >> For my first question. Can you please explain a bit more? I tested with >> two concurrent client requests. I saw two threads running in my service >> routine. My server code does not spawn any thread. So I think gRPC spawns >> the thread for me. That is why I want to understand the thread model. Is it >> one thread per one client? or there is a thread pool for all clients? >> >> Thank you! >> >> Carl Mastrangelo於 2017年1月3日星期二 UTC-8下午4時35分42秒寫道： >>> >>> Each stream does not require a separate thread; gRPC multiplexes these >>> onto your threads. It is slightly different for each language, but in >>> general threading isn't that much of a problem. >>> >>> >>> As for your second question, can you mention what language you are >>> using? That will affect the answer. >>> >>> On Tuesday, January 3, 2017 at 10:27:41 AM UTC-8, [email protected] >>> wrote: >>>> >>>> Hello, >>>> >>>> I am very interested in the gRPC bi-directional streaming feature. I >>>> have a few questions on the server side. In the .proto file, I define a >>>> simple streaming RPC and message below >>>> >>>> service MfMessaging { >>>> rpc MfRpc (stream MfStream) returns (stream MfStream) {} >>>> } >>>> message MfStream { >>>> bytes message=1; >>>> } >>>> >>>> My first question is: what is the thread model at server side? I mean >>>> how gRPC handles if there are multiple concurrent requests from different >>>> clients? Does gRPC spawn a thread for each client? If yes, is there any >>>> limit or gRPC has a thread pool underneath? >>>> >>>> My second question is: In the service routine, I get a >>>> ServerReaderWriter stream (e.g. ServerReaderWriter<MfStream, MfStream>* >>>> stream). What is lifetime of this stream? Can I pass this stream to >>>> another >>>> thread which writes response back later? >>>> >>>> Look forward to your reply. Thank you very much! >>>> >>>> -- You received this message because you are subscribed to the Google Groups "grpc.io" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/grpc-io. To view this discussion on the web visit https://groups.google.com/d/msgid/grpc-io/5e5c01e6-35b7-464a-ad8a-4835f9c13d26%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
