add pcu_name to pcufailed_notice

[monitor.git] / policy.py

diff --git a/policy.py b/policy.py
index edff7bd..cb5d93f 100755 (executable)
--- a/policy.py
+++ b/policy.py
@@ -76,7 +76,7 @@ def main(hostnames, sitenames):
                                # NOTE: there is a narrow window in which this command must be
                                #               evaluated, otherwise the notice will not go out.  
                                #               this is not ideal.
-                               sitehist.sendMessage('online_notice', hostname=host, viart=False)
+                               sitehist.sendMessage('online_notice', hostname=host, viart=False, saveact=True)
                                print "send message for host %s online" % host
 
 
@@ -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')
@@ -134,9 +174,13 @@ def main(hostnames, sitenames):
        for i,site in enumerate(sitenames):
                sitehist = SiteInterface.get_or_make(loginbase=site)
                siteblack = BlacklistRecord.get_by(loginbase=site)
+               skip_due_to_blacklist=False
 
                if siteblack and not siteblack.expired():
                        print "skipping %s due to blacklist.  will expire %s" % (site, siteblack.willExpire() )
+                       skip_due_to_blacklist=True
+                       sitehist.clearPenalty()
+                       sitehist.applyPenalty()
                        continue
 
                # TODO: make query only return records within a certin time range,
@@ -253,6 +297,6 @@ if __name__ == "__main__":
                session.flush()
                sys.exit(0)
        except:
-               #email_exception()
+               email_exception()
                print traceback.print_exc();
                print "fail all..."