tustvold commented on issue #37720:
URL: https://github.com/apache/arrow/issues/37720#issuecomment-1719221381

   > no server side state to clean up (via timeout or other mechanism)
   
   I think this alludes to something I think would make the current protocol 
hard to implement even if you could guarantee that requests are routed to a 
particular server. The server has no reliable mechanism to tell when it can 
clean up after a session, this is because unlike the protocols on which 
FlightSQL appears to be based, **gRPC is not a connection-oriented transport**. 
That is unlike say postgres where a session state can be associated with a 
particular TCP connection, there is no equivalent concept for gRPC. Most (if 
not all) gRPC implementations do not expose connection-level hooks for this 
reason, clients should be free to multiplex as many or as few "sessions" on the 
underlying HTTP/2.0 transport as they deem fit.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to