+# Node states:
+
+def check_node_state(rec, node):
+
+ node_state = rec.observed_status
+ if rec.plc_node_stats:
+ print rec.plc_node_stats
+ boot_state = rec.plc_node_stats['boot_state']
+ last_contact = rec.plc_node_stats['last_contact']
+ node.plc_nodeid = rec.plc_node_stats['node_id']
+ else:
+ boot_state = "unknown"
+ last_contact = None
+
+ if boot_state == 'disable': boot_state = 'disabled'
+ if boot_state == 'diag' or boot_state == 'diagnose': boot_state = 'safeboot'
+
+ if len(rec.plc_node_stats['pcu_ids']) > 0:
+ node.haspcu = True
+ else:
+ node.haspcu = False
+
+
+ # NOTE: 'DOWN' and 'DEBUG' are temporary states, so only need
+ # 'translations' into the node.status state
+ # 'BOOT' is a permanent state, but we want it to have a bit of
+ # hysteresis (less than 0.5 days)
+
+ #################################################################
+ # "Initialize" the findbad states into nodebad status if they are not already set
+
+ if node_state == 'DOWN' and ( node.status != 'offline' and node.status != 'down' ) and boot_state != 'disabled' :
+ print "changed status from %s to offline" % node.status
+ node.status = 'offline'
+ node.last_changed = datetime.now()
+
+ if node_state == 'DEBUG' and node.status != 'monitordebug' and \
+ node.status != 'disabled' and \
+ node.status != 'safeboot':
+ if boot_state != 'disabled' and boot_state != 'safeboot':
+
+ print "changed status from %s to monitordebug" % (node.status)
+ node.status = "monitordebug"
+ node.last_changed = datetime.now()
+ else:
+ print "changed status from %s to %s" % (node.status, boot_state)
+ node.status = boot_state
+ node.last_changed = datetime.now()
+
+ if node_state == 'BOOT' and node.status != 'online' and node.status != 'good':
+ print "changed status from %s to online" % node.status
+ node.status = 'online'
+ node.last_changed = datetime.now()
+
+ #################################################################
+ # Switch temporary hystersis states into their 'firm' states.
+ # online -> good after half a day
+ # offline -> down after two days
+ # monitordebug -> down after 30 days
+ # safeboot -> monitordebug after 60 days
+ # disabled -> down after 60 days
+
+ if node.status == 'online' and changed_greaterthan(node.last_changed, 0.5):
+ print "changed status from %s to good" % node.status
+ node.status = 'good'
+ # NOTE: do not reset last_changed, or you lose how long it's been up.
+
+ if node.status == 'offline' and changed_greaterthan(node.last_changed, 2):
+ print "changed status from %s to down" % node.status
+ node.status = 'down'
+ # NOTE: do not reset last_changed, or you lose how long it's been down.
+
+ if node.status == 'monitordebug' and changed_greaterthan(node.last_changed, 30):
+ print "changed status from %s to down" % node.status
+ node.status = 'down'
+ # NOTE: do not reset last_changed, or you lose how long it's been down.
+
+ if node.status == 'safeboot' and changed_greaterthan(node.last_changed, 60):
+ print "changed status from %s to down" % node.status
+ # NOTE: change an admin mode back into monitordebug after two months.
+ node.status = 'monitordebug'
+ node.last_changed = datetime.now()
+
+ # extreme cases of offline nodes
+ if ( boot_state == 'disabled' or last_contact == None ) and \
+ changed_greaterthan(node.last_changed, 2*30) and \
+ node.status != 'down':
+ print "changed status from %s to down" % node.status
+ node.status = 'down'
+ node.last_changed = datetime.now()
+