1) An HTTP/2 stream is what you are looking for. A stream can be terminated 
with a RST_STREAM frame. Please refer 
https://http2.github.io/http2-spec/#RST_STREAM
2) The RPC would also be terminated. The application would be able to 
retry/restart the RPC if necessary.
3) Yes
4) The RPCs are terminated with the status showing the proper error 
message. 
On Wednesday, February 26, 2020 at 10:37:55 AM UTC-8 [email protected] 
wrote:

> As mentioned in the grpc documentation, a channel is backed by multiple 
> HTTP/2 requests. I was wondering about the behaviour/details in below 
> situations:
> 1) Does a channel have something like a failed HTTP/2 call even if the 
> connection(tcp) is fine ?
> 2) If so, what happens to long running rpcs in such case.
> 3) Is a async streaming rpc served by a single HTTP/2 request over a 
> single tcp connection ?
> 4) As mentioned in docs, a channel may retry to fix transient errors in 
> the tcp connection, so for long running rpcs does this have any impact ?
>
> Thanks
>

-- 
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 view this discussion on the web visit 
https://groups.google.com/d/msgid/grpc-io/c8c24886-f814-40fd-83fb-7978ac5c059a%40googlegroups.com.

Reply via email to