As I found we have written above logic to check availability of one cluster
and assume that all the other clusters have been initialized because of the
Application Clusters created event.
IMO if we are writing such logic we need to add some comments describing
the assumptions. Otherwise it would be difficult for others to understand.
Thanks
On Sat, Feb 7, 2015 at 3:44 PM, Imesh Gunaratne <im...@apache.org> wrote:
> Hi Reka,
>
> Would you mind explaining the logic in this method? It is difficult to
> understand the statements highlighted in yellow. If we are trying to check
> whether all clusters have been initialized in an application, why do we
> return true if one of the service clusters exists?
>
> public static boolean allClustersInitialized(Application application) {
> boolean allClustersInitialized = false;
> for (ClusterDataHolder holder : application.getClusterDataRecursively()) {
> TopologyManager.acquireReadLockForCluster(holder.getServiceType(),
> holder.getClusterId());
>
> try {
> Topology topology = TopologyManager.getTopology();
> if (topology != null) {
> Service service =
> topology.getService(holder.getServiceType());
> if (service != null) {
> if (service.clusterExists(holder.getClusterId())) {
> allClustersInitialized = true;
> return allClustersInitialized;
> } else {
> if (log.isDebugEnabled()) {
> log.debug("[Cluster] " + holder.getClusterId() +
> " is not found in " +
> "the Topology");
> }
> allClustersInitialized = false;
> }
> } else {
> if (log.isDebugEnabled()) {
> log.debug("Service is null in the
> CompleteTopologyEvent");
> }
> }
> } else {
> if (log.isDebugEnabled()) {
> log.debug("Topology is null in the
> CompleteTopologyEvent");
> }
> }
> } finally {
> TopologyManager.releaseReadLockForCluster(holder.getServiceType(),
> holder.getClusterId());
> }
> }
> return allClustersInitialized;
> }
>
> Thanks
>
>
>
> --
> Imesh Gunaratne
>
> Technical Lead, WSO2
> Committer & PMC Member, Apache Stratos
>
--
Imesh Gunaratne
Technical Lead, WSO2
Committer & PMC Member, Apache Stratos
