Thanks Nirmal, that was really valuable information. I couldn't find a member terminated event in the Wiki [1]. This page is incomplete for the most part I guess.
[1] - https://cwiki.apache.org/confluence/display/STRATOS/4.0.0+Stratos+Pub-Sub+Communication On Sat, May 3, 2014 at 2:12 AM, Nirmal Fernando <[email protected]> wrote: > Further if you want, you can build your own object model based on the events > you listen. > > > On Sat, May 3, 2014 at 2:11 AM, Nirmal Fernando <[email protected]> > wrote: >> >> >> >> >> On Sat, May 3, 2014 at 2:01 AM, Akila Ravihansa Perera >> <[email protected]> wrote: >>> >>> Thanks Nirmal, that was very helpful. >>> >>> But what about a member deactivation event as a result of scale-down? >> >> >> It is the same as activation. Autoscaler decides to scale-down.... and >> once member is ready to be shutdown, cloud-controller would terminate the >> member and send out a member terminated event. >>> >>> >>> Just to clarify, if we are going to implement topology event extension >>> points for Cartridge Agent, then it should also process the complete >>> topology and complete tenant events only once according to this >>> pattern? >>> >> >> So, the best practice would be, to listen for the complete* events in >> order to initialize and then identify the events that you would want to >> listen and act only upon those events, thereafter. >>> >>> >>> >>> On Sat, May 3, 2014 at 1:37 AM, Nirmal Fernando <[email protected]> >>> wrote: >>> > Hi Akila, >>> > >>> > Complete* events are there only for the system to withstand a restart. >>> > So, >>> > each of the server would initially wait till it receives the complete* >>> > event >>> > and adjust it state to the current system state and then from that >>> > point >>> > onwards, each server would react on the events that occurs. >>> > >>> > If I take your sample, in the case of member activation during a server >>> > life >>> > time, servers would listen to MemberActivatedEvents and process. >>> > >>> > >>> > On Sat, May 3, 2014 at 1:09 AM, Akila Ravihansa Perera >>> > <[email protected]> >>> > wrote: >>> >> >>> >> Hi Imesh, >>> >> >>> >> I noticed that you have made a commit [1] that will make the LB to >>> >> process complete topology and complete tenant events only once. But >>> >> could you explain the reason for it to be like that? Shouldn't the LB >>> >> be aware of topology changes that might occur in the future? It might >>> >> not be aware of members getting active/inactive dynamically. >>> >> >>> >> I'm just trying to understand the workflow here, would really >>> >> appreciate if anyone can provide some background information on this. >>> >> >>> >> Thanks! >>> >> >>> >> >>> >> [1] >>> >> >>> >> https://github.com/apache/incubator-stratos/commit/783197eaba9edd70212ca70b39679502274fd230 >>> >> >>> >> >>> >> -- >>> >> Akila Ravihansa Perera >>> >> Software Engineer >>> >> WSO2 Inc. >>> >> http://wso2.com >>> >> >>> >> Phone: +94 77 64 154 38 >>> >> Blog: http://ravihansa3000.blogspot.com >>> > >>> > >>> > >>> > >>> > -- >>> > Best Regards, >>> > Nirmal >>> > >>> > Nirmal Fernando. >>> > PPMC Member & Committer of Apache Stratos, >>> > Senior Software Engineer, WSO2 Inc. >>> > >>> > Blog: http://nirmalfdo.blogspot.com/ >>> >>> >>> >>> -- >>> Akila Ravihansa Perera >>> Software Engineer >>> WSO2 Inc. >>> http://wso2.com >>> >>> Phone: +94 77 64 154 38 >>> Blog: http://ravihansa3000.blogspot.com >> >> >> >> >> -- >> Best Regards, >> Nirmal >> >> Nirmal Fernando. >> PPMC Member & Committer of Apache Stratos, >> Senior Software Engineer, WSO2 Inc. >> >> Blog: http://nirmalfdo.blogspot.com/ > > > > > -- > Best Regards, > Nirmal > > Nirmal Fernando. > PPMC Member & Committer of Apache Stratos, > Senior Software Engineer, WSO2 Inc. > > Blog: http://nirmalfdo.blogspot.com/ -- Akila Ravihansa Perera Software Engineer WSO2 Inc. http://wso2.com Phone: +94 77 64 154 38 Blog: http://ravihansa3000.blogspot.com
