Add a mechanism for keeping our vSphere Agents alive

src/vpoller/worker.py

@@ -31,7 +31,7 @@
 import types
 import logging
 import ConfigParser
-from time import asctime
+from time import time, asctime
 
 import zmq
 from vpoller.core import VPollerException
@@ -85,7 +85,10 @@ def run(self, config):
         # Enter the main daemon loop from here
         logging.debug('Entering main daemon loop')
         while not self.time_to_die:
-            socks = dict(self.zpoller.poll())
+            # Keep our vSphere Agents alive
+            self.keep_agents_alive()
+
+            socks = dict(self.zpoller.poll(1000))
 
             # Worker socket, receives client messages for processing
             #
@@ -210,6 +213,7 @@ def spawn_vsphere_agents(self):
                 pwd=each_agent['pwd'],
                 host=each_agent['host']
             )
+            agent.last_keep_alive_heartbeat = time()
             self.agents[agent.host] = agent
 
     def start_vsphere_agents(self):
@@ -222,6 +226,20 @@ def start_vsphere_agents(self):
         for agent in self.agents:
             self.agents[agent].connect()
 
+    def keep_agents_alive(self):
+        """
+        Dummy method to keep our vSphere Agents alive
+        
+        This dummy method calls CurrentTime() vSphere method
+        periodically (every 60 seconds) in order to keep the vSphere Agents alive
+        
+        """
+        for each_agent in self.agents:
+            if (time() - self.agents[each_agent].last_keep_alive_heartbeat) > 60.0:
+                logging.debug('[%s] Agent keep-alive heartbeat', self.agents[each_agent].host)
+                self.agents[each_agent].si.CurrentTime()
+                self.agents[each_agent].last_keep_alive_heartbeat = time()
+
     def shutdown_vsphere_agents(self):
         """
         Disconnects all vPoller Agents from their respective VMware vSphere hosts