Changed the management node querying process to be asynchronous (non-blocking). This is because when nodes are added to the cluster, the management node takes a long time to query kubernetes to get node membership. So if the KVS queries the management node synchronously, it will hang for a long time blocking other user requests.
Added check in case git fetch fails in the start script.
Changed the management node querying process to be asynchronous (non-blocking). This is because when nodes are added to the cluster, the management node takes a long time to query kubernetes to get node membership. So if the KVS queries the management node synchronously, it will hang for a long time blocking other user requests.
Added check in case git fetch fails in the start script.