Correct bug where starting a running instance triggered failed state

pvcd/VMInstance.py

@@ -115,6 +115,19 @@ class VMInstance:
             self.instart = False
             return
    
+        # Try to get the current state in case it's already running
+        try:
+            self.dom = self.lookupByUUID(self.domuuid)
+            curstate = self.dom.state()[0]
+        except:
+            curstate = 'notstart'
+
+        if curstate == libvirt.VIR_DOMAIN_RUNNING:
+            # If it is running just update the model
+            self.addDomainToList()
+            zkhandler.writedata(self.zk_conn, { '/domains/{}/failedreason'.format(self.domuuid): '' })
+        else:
+            # Or try to create it
             try:
                 # Grab the domain information from Zookeeper
                 xmlconfig = zkhandler.readdata(self.zk_conn, '/domains/{}/xml'.format(self.domuuid))