Hi All, As from earlier model, we were wiping out the monitors when we undeploy the application. But we are currently starting the monitors when we deploy the deployment policy. So, when undeploying the policy only, we will have to wipe out all the clusters and the monitors. In that case, the flow for the undeployment of the policy will be as below:
Undeployment of policy ------------------------------ - Update the application monitor with the flag saying terminating=true - All Application instances will get updated as state = terminating - Underlying groups instances and cluster instances will be updated as state = terminating by the parent notification - Upon cluster instance terminating event, all the members will get to wiped out in the cluster instance - Cluster Instance will get to Terminated once all the members are wiped out - Once all the clusters instances are wiped out, cluster monitor will get wiped out - Upon cluster Instance terminated, group instance and then application instance will be terminated - Once all the application instances are wiped out, app monitor will get to removed and send application Terminated event. - Call CC to remove all the clusters related to application in Topology - Autosclaer will remove the in-memory clusters upon Topology ClustersRemovedEvent Please share your concerns whether we can go ahead with this flow. Thanks, Reka -- Reka Thirunavukkarasu Senior Software Engineer, WSO2, Inc.:http://wso2.com, Mobile: +94776442007
