X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=policy.py;h=cb5d93f31d52c3700296fa95b09cc751aa47c563;hb=e2e493681014a07579fee2719a683b364a72fe0a;hp=84bdb445a17e40086f1f07ff4f4eca6ea4f149ce;hpb=cb01884d466b6250c1f7351e3b68c7c8a3699c62;p=monitor.git

diff --git a/policy.py b/policy.py
index 84bdb44..cb5d93f 100755
--- a/policy.py
+++ b/policy.py
@@ -91,25 +91,60 @@ def main(hostnames, sitenames):
 		# if it is offline and HAS a PCU, then try to use it.
 		if nodehist.haspcu and nodehist.status in ['offline', 'down'] and \
 			changed_greaterthan(nodehist.last_changed,1.0) and \
+			not nodehist.firewall and \
 			not found_between(recent_actions, 'try_reboot', 3.5, 1):
 
+				# TODO: there MUST be a better way to do this... 
+				# get fb node record for pcuid
+				fbpcu = None
+				fbnode = FindbadNodeRecord.get_latest_by(hostname=host)
+				if fbnode:
+					fbpcu = FindbadPCURecord.get_latest_by(plc_pcuid=fbnode.plc_pcuid)
+
 				sitehist.attemptReboot(host)
 				print "send message for host %s try_reboot" % host
+				if not fbpcu.test_is_ok() and \
+					not found_within(recent_actions, 'pcuerror_notice', 3.0):
+
+					args = {}
+					if fbpcu:
+						args['pcu_name'] = fbpcu.pcu_name()
+						args['pcu_errors'] = fbpcu.pcu_errors()
+					else:
+						args['pcu_name'] = "error looking up pcu name"
+						args['pcu_errors'] = ""
+
+					args['hostname'] = host
+					sitehist.sendMessage('pcuerror_notice', **args)
+					print "send message for host %s PCU Failure" % host
+					
 
 		# NOTE: non-intuitive is that found_between(try_reboot, 3.5, 1)
 		# 		will be false for a day after the above condition is satisfied
 		if nodehist.haspcu and nodehist.status in ['offline', 'down'] and \
 			changed_greaterthan(nodehist.last_changed,1.5) and \
+			not nodehist.firewall and \
 			found_between(recent_actions, 'try_reboot', 3.5, 1) and \
 			not found_within(recent_actions, 'pcufailed_notice', 3.5):
 				
+				# TODO: there MUST be a better way to do this... 
+				# get fb node record for pcuid
+				fbpcu = None
+				fbnode = FindbadNodeRecord.get_latest_by(hostname=host)
+				if fbnode:
+					fbpcu = FindbadPCURecord.get_latest_by(plc_pcuid=fbnode.plc_pcuid)
+				if fbpcu:
+					pcu_name = fbpcu.pcu_name()
+				else:
+					pcu_name = "error looking up pcu name"
+
+				# get fb pcu record for pcuid
 				# send pcu failure message
-				#act = ActionRecord(**kwargs)
-				sitehist.sendMessage('pcufailed_notice', hostname=host)
+				sitehist.sendMessage('pcufailed_notice', hostname=host, pcu_name=pcu_name)
 				print "send message for host %s PCU Failure" % host
 
-		if nodehist.status == 'monitordebug' and \
-			changed_greaterthan(nodehist.last_changed, 1) and \
+		if nodehist.status == 'failboot' and \
+			changed_greaterthan(nodehist.last_changed, 0.25) and \
 			not found_between(recent_actions, 'bootmanager_restore', 0.5, 0):
 				# send down node notice
 				# delay 0.5 days before retrying...
@@ -119,12 +154,17 @@ def main(hostnames, sitenames):
 			#	sitehist.sendMessage('retry_bootman', hostname=host)
 
 		if nodehist.status == 'down' and \
-			changed_greaterthan(nodehist.last_changed, 2) and \
-			not found_within(recent_actions, 'down_notice', 3.5):
-				# send down node notice
-
-				sitehist.sendMessage('down_notice', hostname=host)
-				print "send message for host %s down" % host
+			changed_greaterthan(nodehist.last_changed, 2):
+				if not nodehist.firewall and not found_within(recent_actions, 'down_notice', 3.5):
+					# send down node notice
+					sitehist.sendMessage('down_notice', hostname=host)
+					print "send message for host %s down" % host
+
+				if nodehist.firewall and not found_within(recent_actions, 'firewall_notice', 3.5):
+					# send down node notice
+					#email_exception(host, "firewall_notice")
+					sitehist.sendMessage('firewall_notice', hostname=host)
+					print "send message for host %s down" % host
 
 		node_count = node_count + 1
 		print "time: ", time.strftime('%Y-%m-%d %H:%M:%S')