Re: [Dev] Issue in accessing same boss and worker group in HTTP carbon transport from MSF4J and Siddhi together

2017-08-02 Thread Yasara Dissanayake
Hi Shafreen, Thank you for the information. Regards, Yasara. On Thu, Aug 3, 2017 at 11:02 AM, Shafreen Anfar wrote: > Hi Yasara, > > As part of the restructuring [1] we will be providing a way to create > server connectors with separate event loop groups. > > [1] "Restructuring CarbonMessagePr

Re: [Dev] Issue in accessing same boss and worker group in HTTP carbon transport from MSF4J and Siddhi together

2017-08-02 Thread Shafreen Anfar
Hi Yasara, As part of the restructuring [1] we will be providing a way to create server connectors with separate event loop groups. [1] "Restructuring CarbonMessageProcessor, CarbonMessage and CarbonCallBack" On Tue, Aug 1, 2017 at 1:34 AM, Yasara Dissanayake wrote: > Hi Shfreen, > > It is wor

Re: [Dev] Issue in accessing same boss and worker group in HTTP carbon transport from MSF4J and Siddhi together

2017-07-31 Thread Yasara Dissanayake
Hi Shfreen, It is workable without shutting down EventLoop Groups from ServerConnectorController. But in this scenario AFAIU both MSF4j and Siddhi uses same ServerConnectorController and it gives following issues when considering the use case for siddhi. 1. Cannot configure event group si

Re: [Dev] Issue in accessing same boss and worker group in HTTP carbon transport from MSF4J and Siddhi together

2017-07-31 Thread Shafreen Anfar
Hi Yasara, Why do you want to shutdown the EventLoop groups. AFAIR, it was due to some blocking operation in worker (child) thread pool right ? On Mon, Jul 31, 2017 at 10:49 PM, Yasara Dissanayake wrote: > Hi all, > > > Currently I’m working on implementing HTTP carbon transport for Siddhi. > T

[Dev] Issue in accessing same boss and worker group in HTTP carbon transport from MSF4J and Siddhi together

2017-07-31 Thread Yasara Dissanayake
Hi all, Currently I’m working on implementing HTTP carbon transport for Siddhi. The problem I face there is that since MSF4J also using HTTP carbon transport we are facing the issue of keeping single bossGroup and workGroup for both MSF4J and Siddhi [1]. Therefore we cannot independently control