+= architecture-group Thanks Ramindu for the explanation!
Why can't we just use the member IPs discovered from the membership scheme here? AFAIU what we are doing here is creating a point to point connection here. Therefore, without explicitly specifying the IP address in the eventSync configuration, we can simply use the local member's IP from the Hz instance. Please correct me if I'm wrong. On Wed, Jul 6, 2016 at 8:16 PM, Ramindu De Silva <[email protected]> wrote: > Hi Isuru, > > Please find the explanations inline. > > In a two node HA cluster, in the eventSync/hostName configuration > (event-processor.xml), do we need to provide the own host IP? If so, how > does each nodes discover each other for event synchronization purpose? For > what exact purpose this eventSync configuration is used? > > The provided host IP will be saved in the Hazelcast map after getting the > server's state (passive [1] / active [2] ). So the other members which are > in the cluster can get the relevant configuration using the passive key and > active key. > > > If we add a presenter node in addition to the two worker nodes, I assume > we should point to the presenter node from the both worker nodes. As per > the configuration, seems this is a TCP level connection. Why do we need a > TCP connection to the presenter? > > In, HA configured serves, each and every server contains the same > artifacts, even the presenter node. But the presenter node only publishes > the output which has been already processed by the active node which is > received using the tcp connection [3]. > > FYI, the active node also sends the events to the passing node using the > tcp connection, and the events will be processed in parallel but the > passive node will drop the events from publishing or sending to the > presenter node (if available). > > 1. > https://github.com/wso2/carbon-analytics-common/blob/master/components/event-processor-manager/org.wso2.carbon.event.processor.manager.core/src/main/java/org/wso2/carbon/event/processor/manager/core/internal/HAManager.java#L259 > > > 2. > https://github.com/wso2/carbon-analytics-common/blob/master/components/event-processor-manager/org.wso2.carbon.event.processor.manager.core/src/main/java/org/wso2/carbon/event/processor/manager/core/internal/HAManager.java#L233 > > 3. > https://github.com/wso2/carbon-analytics-common/blob/master/components/event-processor-manager/org.wso2.carbon.event.processor.manager.core/src/main/java/org/wso2/carbon/event/processor/manager/core/internal/EventHandler.java#L161 > > > Best Regards, > > On Wed, Jul 6, 2016 at 4:49 PM, Isuru Haththotuwa <[email protected]> wrote: > >> += wso2-dev list >> <https://docs.wso2.com/display/CLUSTER44x/Clustering+CEP+4.0.0#ClusteringCEP4.0.0-Highavailabilitydeployment> >> >> On Wed, Jul 6, 2016 at 4:34 PM, Isuru Haththotuwa <[email protected]> >> wrote: >> >>> Hi CEP team, >>> >>> I'm going through [1] to understand the CEP HA deployment. Have a few >>> doubts to get clarified: >>> >>> >>> 1. In a two node HA cluster, in the eventSync/hostName configuration >>> (event-processor.xml), do we need to provide the own host IP? If so, how >>> does each nodes discover each other for event synchronization purpose? >>> For >>> what exact purpose this eventSync configuration is used? >>> 2. If we add a presenter node in addition to the two worker nodes, I >>> assume we should point to the presenter node from the both worker nodes. >>> As >>> per the configuration, seems this is a TCP level connection. Why do we >>> need >>> a TCP connection to the presenter? >>> >>> Missed the link: >> >> [1]. >> https://docs.wso2.com/display/CLUSTER44x/Clustering+CEP+4.0.0#ClusteringCEP4.0.0-Highavailabilitydeployment >> >>> >>> -- >>> Thanks and Regards, >>> >>> Isuru H. >>> +94 716 358 048* <http://wso2.com/>* >>> >>> >>> >> >> >> -- >> Thanks and Regards, >> >> Isuru H. >> +94 716 358 048* <http://wso2.com/>* >> >> >> >> _______________________________________________ >> Dev mailing list >> [email protected] >> http://wso2.org/cgi-bin/mailman/listinfo/dev >> >> > > > -- > *Ramindu De Silva* > Software Engineer > WSO2 Inc.: http://wso2.com > lean.enterprise.middleware > > email: [email protected] <[email protected]> > mob: +94 772339350 > mob: +94 719678895 > -- Thanks and Regards, Isuru H. +94 716 358 048* <http://wso2.com/>*
_______________________________________________ Architecture mailing list [email protected] https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture
