When receiving an HTTP(S) connection from a WS client, the request is put into a queue at TCP level of the HTTP(S) server. Next, the incoming requests of the queue are accepted before to be processed. If a HTTP(S) connection arrives when the queue is full, the HTTP(S) connection is refused.
To avoid connection rejection, the queue should be sized according to the expected load. Such a parameter exists at SocketServer level and it is already used by the Petals BC SOAP but it is set with the HTTP acceptor number that is the number of threads that accept HTTP connections.
So the new parameters of PETALSBCSOAP-179 are required before to fix the value to used as queue size.
Bug
Resolved
Major
When receiving an HTTP(S) connection from a WS client, the request is put into a queue at TCP level of the HTTP(S) server. Next, the incoming requests of the queue are accepted before to be processed. If a HTTP(S) connection arrives when the queue is full, the HTTP(S) connection is refused.
To avoid connection rejection, the queue should be sized according to the expected load. Such a parameter exists at SocketServer level and it is already used by the Petals BC SOAP but it is set with the HTTP acceptor number that is the number of threads that accept HTTP connections.
So the new parameters of
PETALSBCSOAP-179are required before to fix the value to used as queue size.PETALSBCSOAP-179are required before to fix the value to used as queue size.