add pcu_name to pcufailed_notice
[monitor.git] / policy.py
index 77cf76e..cb5d93f 100755 (executable)
--- a/policy.py
+++ b/policy.py
@@ -94,8 +94,30 @@ def main(hostnames, sitenames):
                        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
@@ -105,13 +127,24 @@ def main(hostnames, sitenames):
                        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 == 'failboot' and \
-                       changed_greaterthan(nodehist.last_changed, 1) 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...