Source code for shinken.objects.config

#!/usr/bin/python

# -*- coding: utf-8 -*-

# Copyright (C) 2009-2012:
#    Gabes Jean, naparuba@gmail.com
#    Gerhard Lausser, Gerhard.Lausser@consol.de
#    Gregory Starck, g.starck@gmail.com
#    Hartmut Goebel, h.goebel@goebel-consult.de
#
# This file is part of Shinken.
#
# Shinken is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# Shinken is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with Shinken.  If not, see <http://www.gnu.org/licenses/>.

""" Config is the class to read, load and manipulate the user
 configuration. It read a main cfg (nagios.cfg) and get all informations
 from it. It create objects, make link between them, clean them, and cut
 them into independent parts. The main user of this is Arbiter, but schedulers
 use it too (but far less)"""

import re
import sys
import string
import copy
import os
import socket
import itertools
import time
import random
import cPickle
from StringIO import StringIO
from multiprocessing import Process, Manager

from item import Item
from timeperiod import Timeperiod, Timeperiods
from service import Service, Services
from command import Command, Commands
from resultmodulation import Resultmodulation, Resultmodulations
from businessimpactmodulation import Businessimpactmodulation, Businessimpactmodulations
from escalation import Escalation, Escalations
from serviceescalation import Serviceescalation, Serviceescalations
from hostescalation import Hostescalation, Hostescalations
from host import Host, Hosts
from hostgroup import Hostgroup, Hostgroups
from realm import Realm, Realms
from contact import Contact, Contacts
from contactgroup import Contactgroup, Contactgroups
from notificationway import NotificationWay, NotificationWays
from checkmodulation import CheckModulation, CheckModulations
from macromodulation import MacroModulation, MacroModulations
from servicegroup import Servicegroup, Servicegroups
from servicedependency import Servicedependency, Servicedependencies
from hostdependency import Hostdependency, Hostdependencies
from module import Module, Modules
from discoveryrule import Discoveryrule, Discoveryrules
from discoveryrun import Discoveryrun, Discoveryruns
from hostextinfo import HostExtInfo, HostsExtInfo
from serviceextinfo import ServiceExtInfo, ServicesExtInfo
from trigger import Trigger, Triggers
from pack import Pack, Packs

from shinken.util import split_semicolon
from shinken.arbiterlink import ArbiterLink, ArbiterLinks
from shinken.schedulerlink import SchedulerLink, SchedulerLinks
from shinken.reactionnerlink import ReactionnerLink, ReactionnerLinks
from shinken.brokerlink import BrokerLink, BrokerLinks
from shinken.receiverlink import ReceiverLink, ReceiverLinks
from shinken.pollerlink import PollerLink, PollerLinks
from shinken.graph import Graph
from shinken.log import logger, console_logger
from shinken.property import UnusedProp, BoolProp, IntegerProp, CharProp, StringProp, LogLevelProp, ListProp
from shinken.daemon import get_cur_user, get_cur_group

no_longer_used_txt = 'This parameter is not longer take from the main file, but must be defined in the status_dat broker module instead. But Shinken will create you one if there are no present and use this parameter in it, so no worry.'
not_interresting_txt = 'We do not think such an option is interesting to manage.'


[docs]class Config(Item): cache_path = "objects.cache" my_type = "config" # Properties: # *required: if True, there is not default, and the config must put them # *default: if not set, take this value # *pythonize: function call to # *class_inherit: (Service, 'blabla'): must set this property to the # Service class with name blabla # if (Service, None): must set this property to the Service class with # same name # *unused: just to warn the user that the option he use is no more used # in Shinken # *usage_text: if present, will print it to explain why it's no more useful properties = { 'prefix': StringProp(default='/usr/local/shinken/'), 'workdir': StringProp(default='/var/run/shinken/'), 'config_base_dir': StringProp(default=''), # will be set when we will load a file 'modules_dir': StringProp(default='/var/lib/shinken/modules'), 'use_local_log': BoolProp(default='1'), 'log_level': LogLevelProp(default='WARNING'), 'local_log': StringProp(default='/var/log/shinken/arbiterd.log'), 'log_file': UnusedProp(text=no_longer_used_txt), 'object_cache_file': UnusedProp(text=no_longer_used_txt), 'precached_object_file': UnusedProp(text='Shinken does not use precached_object_files. Skipping.'), 'resource_file': StringProp(default='/tmp/resources.txt'), 'temp_file': UnusedProp(text='Temporary files are not used in the shinken architecture. Skipping'), 'status_file': UnusedProp(text=no_longer_used_txt), 'status_update_interval': UnusedProp(text=no_longer_used_txt), 'shinken_user': StringProp(default=get_cur_user()), 'shinken_group': StringProp(default=get_cur_group()), 'enable_notifications': BoolProp(default='1', class_inherit=[(Host, None), (Service, None), (Contact, None)]), 'execute_service_checks': BoolProp(default='1', class_inherit=[(Service, 'execute_checks')]), 'accept_passive_service_checks': BoolProp(default='1', class_inherit=[(Service, 'accept_passive_checks')]), 'execute_host_checks': BoolProp(default='1', class_inherit=[(Host, 'execute_checks')]), 'accept_passive_host_checks': BoolProp(default='1', class_inherit=[(Host, 'accept_passive_checks')]), 'enable_event_handlers': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'log_rotation_method': CharProp(default='d'), 'log_archive_path': StringProp(default='/usr/local/shinken/var/archives'), 'check_external_commands': BoolProp(default='1'), 'command_check_interval': UnusedProp(text='another value than look always the file is useless, so we fix it.'), 'command_file': StringProp(default=''), 'external_command_buffer_slots': UnusedProp(text='We do not limit the external command slot.'), 'check_for_updates': UnusedProp(text='network administrators will never allow such communication between server and the external world. Use your distribution packet manager to know if updates are available or go to the http://www.shinken-monitoring.org website instead.'), 'bare_update_checks': UnusedProp(text=None), 'lock_file': StringProp(default='/var/run/shinken/arbiterd.pid'), 'retain_state_information': UnusedProp(text='sorry, retain state information will not be implemented because it is useless.'), 'state_retention_file': StringProp(default=''), 'retention_update_interval': IntegerProp(default='60'), 'use_retained_program_state': UnusedProp(text=not_interresting_txt), 'use_retained_scheduling_info': UnusedProp(text=not_interresting_txt), 'retained_host_attribute_mask': UnusedProp(text=not_interresting_txt), 'retained_service_attribute_mask': UnusedProp(text=not_interresting_txt), 'retained_process_host_attribute_mask': UnusedProp(text=not_interresting_txt), 'retained_process_service_attribute_mask': UnusedProp(text=not_interresting_txt), 'retained_contact_host_attribute_mask': UnusedProp(text=not_interresting_txt), 'retained_contact_service_attribute_mask': UnusedProp(text=not_interresting_txt), 'use_syslog': BoolProp(default='0'), 'log_notifications': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'log_service_retries': BoolProp(default='1', class_inherit=[(Service, 'log_retries')]), 'log_host_retries': BoolProp(default='1', class_inherit=[(Host, 'log_retries')]), 'log_event_handlers': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'log_initial_states': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'log_external_commands': BoolProp(default='1'), 'log_passive_checks': BoolProp(default='1'), 'global_host_event_handler': StringProp(default='', class_inherit=[(Host, 'global_event_handler')]), 'global_service_event_handler': StringProp(default='', class_inherit=[(Service, 'global_event_handler')]), 'sleep_time': UnusedProp(text='this deprecated option is useless in the shinken way of doing.'), 'service_inter_check_delay_method': UnusedProp(text='This option is useless in the Shinken scheduling. The only way is the smart way.'), 'max_service_check_spread': IntegerProp(default='30', class_inherit=[(Service, 'max_check_spread')]), 'service_interleave_factor': UnusedProp(text='This option is useless in the Shinken scheduling because it use a random distribution for initial checks.'), 'max_concurrent_checks': UnusedProp(text='Limiting the max concurrent checks is not helpful to got a good running monitoring server.'), 'check_result_reaper_frequency': UnusedProp(text='Shinken do not use reaper process.'), 'max_check_result_reaper_time': UnusedProp(text='Shinken do not use reaper process.'), 'check_result_path': UnusedProp(text='Shinken use in memory returns, not check results on flat file.'), 'max_check_result_file_age': UnusedProp(text='Shinken do not use flat file check resultfiles.'), 'host_inter_check_delay_method': UnusedProp(text='This option is unused in the Shinken scheduling because distribution of the initial check is a random one.'), 'max_host_check_spread': IntegerProp(default='30', class_inherit=[(Host, 'max_check_spread')]), 'interval_length': IntegerProp(default='60', class_inherit=[(Host, None), (Service, None)]), 'auto_reschedule_checks': BoolProp(managed=False, default='1'), 'auto_rescheduling_interval': IntegerProp(managed=False, default='1'), 'auto_rescheduling_window': IntegerProp(managed=False, default='180'), 'use_aggressive_host_checking': BoolProp(default='0', class_inherit=[(Host, None)]), 'translate_passive_host_checks': BoolProp(managed=False, default='1'), 'passive_host_checks_are_soft': BoolProp(managed=False, default='1'), 'enable_predictive_host_dependency_checks': BoolProp(managed=False, default='1', class_inherit=[(Host, 'enable_predictive_dependency_checks')]), 'enable_predictive_service_dependency_checks': StringProp(managed=False, default='1'), 'cached_host_check_horizon': IntegerProp(default='0', class_inherit=[(Host, 'cached_check_horizon')]), 'cached_service_check_horizon': IntegerProp(default='0', class_inherit=[(Service, 'cached_check_horizon')]), 'use_large_installation_tweaks': UnusedProp(text='this option is deprecated because in shinken it is just an alias for enable_environment_macros=0'), 'free_child_process_memory': UnusedProp(text='this option is automatic in Python processes'), 'child_processes_fork_twice': UnusedProp(text='fork twice is not use.'), 'enable_environment_macros': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'enable_flap_detection': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'low_service_flap_threshold': IntegerProp(default='20', class_inherit=[(Service, 'global_low_flap_threshold')]), 'high_service_flap_threshold': IntegerProp(default='30', class_inherit=[(Service, 'global_high_flap_threshold')]), 'low_host_flap_threshold': IntegerProp(default='20', class_inherit=[(Host, 'global_low_flap_threshold')]), 'high_host_flap_threshold': IntegerProp(default='30', class_inherit=[(Host, 'global_high_flap_threshold')]), 'soft_state_dependencies': BoolProp(managed=False, default='0'), 'service_check_timeout': IntegerProp(default='60', class_inherit=[(Service, 'check_timeout')]), 'host_check_timeout': IntegerProp(default='30', class_inherit=[(Host, 'check_timeout')]), 'timeout_exit_status': IntegerProp(default='2'), 'event_handler_timeout': IntegerProp(default='30', class_inherit=[(Host, None), (Service, None)]), 'notification_timeout': IntegerProp(default='30', class_inherit=[(Host, None), (Service, None)]), 'ocsp_timeout': IntegerProp(default='15', class_inherit=[(Service, None)]), 'ochp_timeout': IntegerProp(default='15', class_inherit=[(Host, None)]), 'perfdata_timeout': IntegerProp(default='5', class_inherit=[(Host, None), (Service, None)]), 'obsess_over_services': BoolProp(default='0', class_inherit=[(Service, 'obsess_over')]), 'ocsp_command': StringProp(default='', class_inherit=[(Service, None)]), 'obsess_over_hosts': BoolProp(default='0', class_inherit=[(Host, 'obsess_over')]), 'ochp_command': StringProp(default='', class_inherit=[(Host, None)]), 'process_performance_data': BoolProp(default='1', class_inherit=[(Host, None), (Service, None)]), 'host_perfdata_command': StringProp(default='', class_inherit=[(Host, 'perfdata_command')]), 'service_perfdata_command': StringProp(default='', class_inherit=[(Service, 'perfdata_command')]), 'host_perfdata_file': StringProp(default='', class_inherit=[(Host, 'perfdata_file')]), 'service_perfdata_file': StringProp(default='', class_inherit=[(Service, 'perfdata_file')]), 'host_perfdata_file_template': StringProp(default='/tmp/host.perf', class_inherit=[(Host, 'perfdata_file_template')]), 'service_perfdata_file_template': StringProp(default='/tmp/host.perf', class_inherit=[(Service, 'perfdata_file_template')]), 'host_perfdata_file_mode': CharProp(default='a', class_inherit=[(Host, 'perfdata_file_mode')]), 'service_perfdata_file_mode': CharProp(default='a', class_inherit=[(Service, 'perfdata_file_mode')]), 'host_perfdata_file_processing_interval': IntegerProp(managed=False, default='15'), 'service_perfdata_file_processing_interval': IntegerProp(managed=False, default='15'), 'host_perfdata_file_processing_command': StringProp(managed=False, default='', class_inherit=[(Host, 'perfdata_file_processing_command')]), 'service_perfdata_file_processing_command': StringProp(managed=False, default=None), 'check_for_orphaned_services': BoolProp(default='1', class_inherit=[(Service, 'check_for_orphaned')]), 'check_for_orphaned_hosts': BoolProp(default='1', class_inherit=[(Host, 'check_for_orphaned')]), 'check_service_freshness': BoolProp(default='1', class_inherit=[(Service, 'global_check_freshness')]), 'service_freshness_check_interval': IntegerProp(default='60'), 'check_host_freshness': BoolProp(default='1', class_inherit=[(Host, 'global_check_freshness')]), 'host_freshness_check_interval': IntegerProp(default='60'), 'additional_freshness_latency': IntegerProp(default='15', class_inherit=[(Host, None), (Service, None)]), 'enable_embedded_perl': BoolProp(managed=False, default='1', help='It will surely never be managed, but it should not be useful with poller performances.'), 'use_embedded_perl_implicitly': BoolProp(managed=False, default='0'), 'date_format': StringProp(managed=False, default=None), 'use_timezone': StringProp(default='', class_inherit=[(Host, None), (Service, None), (Contact, None)]), 'illegal_object_name_chars': StringProp(default="""`~!$%^&*"|'<>?,()=""", class_inherit=[(Host, None), (Service, None), (Contact, None), (HostExtInfo, None)]), 'illegal_macro_output_chars': StringProp(default='', class_inherit=[(Host, None), (Service, None), (Contact, None)]), 'use_regexp_matching': BoolProp(managed=False, default='0', help=' if you go some host or service definition like prod*, it will surely failed from now, sorry.'), 'use_true_regexp_matching': BoolProp(managed=False, default=None), 'admin_email': UnusedProp(text='sorry, not yet implemented.'), 'admin_pager': UnusedProp(text='sorry, not yet implemented.'), 'event_broker_options': UnusedProp(text='event broker are replaced by modules with a real configuration template.'), 'broker_module': StringProp(default=''), 'debug_file': UnusedProp(text=None), 'debug_level': UnusedProp(text=None), 'debug_verbosity': UnusedProp(text=None), 'max_debug_file_size': UnusedProp(text=None), 'modified_attributes': IntegerProp(default=0L), #'$USERn$: {'required':False, 'default':''} # Add at run in __init__ # SHINKEN SPECIFIC 'idontcareaboutsecurity': BoolProp(default='0'), 'daemon_enabled' : BoolProp(default='1'), # Put to 0 to disable the arbiter to run 'daemon_thread_pool_size' : IntegerProp(default='8'), 'flap_history': IntegerProp(default='20', class_inherit=[(Host, None), (Service, None)]), 'max_plugins_output_length': IntegerProp(default='8192', class_inherit=[(Host, None), (Service, None)]), 'no_event_handlers_during_downtimes': BoolProp(default='0', class_inherit=[(Host, None), (Service, None)]), # Interval between cleaning queues pass 'cleaning_queues_interval': IntegerProp(default='900'), # Enable or not the notice about old Nagios parameters 'disable_old_nagios_parameters_whining': BoolProp(default='0'), # Now for problem/impact states changes 'enable_problem_impacts_states_change': BoolProp(default='0', class_inherit=[(Host, None), (Service, None)]), # More a running value in fact 'resource_macros_names': ListProp(default=[]), # SSL PART # global boolean for know if we use ssl or not 'use_ssl': BoolProp(default='0', class_inherit=[(SchedulerLink, None), (ReactionnerLink, None), (BrokerLink, None), (PollerLink, None), \ (ReceiverLink, None), (ArbiterLink, None)]), 'ca_cert': StringProp(default='etc/certs/ca.pem'), 'server_cert': StringProp(default='etc/certs/server.cert'), 'server_key': StringProp(default='etc/certs/server.key'), 'hard_ssl_name_check': BoolProp(default='0'), # Log format 'human_timestamp_log': BoolProp(default='0'), ## Discovery part 'strip_idname_fqdn': BoolProp(default='1'), 'runners_timeout': IntegerProp(default='3600'), # pack_distribution_file is for keeping a distribution history # of the host distribution in the several "packs" so a same # scheduler will have more change of getting the same host 'pack_distribution_file': StringProp(default='pack_distribution.dat'), ## WEBUI part 'webui_lock_file': StringProp(default='webui.pid'), 'webui_port': IntegerProp(default='8080'), 'webui_host': StringProp(default='0.0.0.0'), # Large env tweacks 'use_multiprocesses_serializer': BoolProp(default='0'), } macros = { 'PREFIX': 'prefix', 'MAINCONFIGFILE': '', 'STATUSDATAFILE': '', 'COMMENTDATAFILE': '', 'DOWNTIMEDATAFILE': '', 'RETENTIONDATAFILE': '', 'OBJECTCACHEFILE': '', 'TEMPFILE': '', 'TEMPPATH': '', 'LOGFILE': '', 'RESOURCEFILE': '', 'COMMANDFILE': 'command_file', 'HOSTPERFDATAFILE': '', 'SERVICEPERFDATAFILE': '', 'ADMINEMAIL': '', 'ADMINPAGER': '' #'USERn': '$USERn$' # Add at run time } # We create dict of objects # Type: 'name in objects': {Class of object, Class of objects, # 'property for self for the objects(config)' types_creations = { 'timeperiod': (Timeperiod, Timeperiods, 'timeperiods'), 'service': (Service, Services, 'services'), 'servicegroup': (Servicegroup, Servicegroups, 'servicegroups'), 'command': (Command, Commands, 'commands'), 'host': (Host, Hosts, 'hosts'), 'hostgroup': (Hostgroup, Hostgroups, 'hostgroups'), 'contact': (Contact, Contacts, 'contacts'), 'contactgroup': (Contactgroup, Contactgroups, 'contactgroups'), 'notificationway': (NotificationWay, NotificationWays, 'notificationways'), 'checkmodulation': (CheckModulation, CheckModulations, 'checkmodulations'), 'macromodulation': (MacroModulation, MacroModulations, 'macromodulations'), 'servicedependency': (Servicedependency, Servicedependencies, 'servicedependencies'), 'hostdependency': (Hostdependency, Hostdependencies, 'hostdependencies'), 'arbiter': (ArbiterLink, ArbiterLinks, 'arbiters'), 'scheduler': (SchedulerLink, SchedulerLinks, 'schedulers'), 'reactionner': (ReactionnerLink, ReactionnerLinks, 'reactionners'), 'broker': (BrokerLink, BrokerLinks, 'brokers'), 'receiver': (ReceiverLink, ReceiverLinks, 'receivers'), 'poller': (PollerLink, PollerLinks, 'pollers'), 'realm': (Realm, Realms, 'realms'), 'module': (Module, Modules, 'modules'), 'resultmodulation': (Resultmodulation, Resultmodulations, 'resultmodulations'), 'businessimpactmodulation': (Businessimpactmodulation, Businessimpactmodulations, 'businessimpactmodulations'), 'escalation': (Escalation, Escalations, 'escalations'), 'serviceescalation': (Serviceescalation, Serviceescalations, 'serviceescalations'), 'hostescalation': (Hostescalation, Hostescalations, 'hostescalations'), 'discoveryrule': (Discoveryrule, Discoveryrules, 'discoveryrules'), 'discoveryrun': (Discoveryrun, Discoveryruns, 'discoveryruns'), 'hostextinfo': (HostExtInfo, HostsExtInfo, 'hostsextinfo'), 'serviceextinfo': (ServiceExtInfo, ServicesExtInfo, 'servicesextinfo'), } # This tab is used to transform old parameters name into new ones # so from Nagios2 format, to Nagios3 ones old_properties = { 'nagios_user': 'shinken_user', 'nagios_group': 'shinken_group', 'modulesdir': 'modules_dir', } read_config_silent = 0 early_created_types = ['arbiter', 'module'] configuration_types = ['void', 'timeperiod', 'command', 'contactgroup', 'hostgroup', 'contact', 'notificationway', 'checkmodulation', 'macromodulation', 'host', 'service', 'servicegroup', 'servicedependency', 'hostdependency', 'arbiter', 'scheduler', 'reactionner', 'broker', 'receiver', 'poller', 'realm', 'module', 'resultmodulation', 'escalation', 'serviceescalation', 'hostescalation', 'discoveryrun', 'discoveryrule', 'businessimpactmodulation', 'hostextinfo', 'serviceextinfo'] def __init__(self): self.params = {} self.resource_macros_names = [] # By default the conf is correct self.conf_is_correct = True # We tag the conf with a magic_hash, a random value to # idify this conf random.seed(time.time()) self.magic_hash = random.randint(1, 100000) self.configuration_errors = [] self.triggers_dirs = [] self.triggers = Triggers({}) self.packs_dirs = [] self.packs = Packs({})
[docs] def get_name(self): return 'global configuration file' # We've got macro in the resource file and we want # to update our MACRO dict with it
[docs] def fill_resource_macros_names_macros(self): """ fill the macro dict will all value from self.resource_macros_names""" properties = self.__class__.properties macros = self.__class__.macros for macro_name in self.resource_macros_names: properties['$' + macro_name + '$'] = StringProp(default='') macros[macro_name] = '$' + macro_name + '$'
[docs] def load_params(self, params): for elt in params: elts = elt.split('=', 1) if len(elts) == 1: # error, there is no = ! self.conf_is_correct = False logger.error("[config] the parameter %s is malformed! (no = sign)" % elts[0]) else: self.params[elts[0]] = elts[1] setattr(self, elts[0], elts[1]) # Maybe it's a variable as $USER$ or $ANOTHERVATRIABLE$ # so look at the first character. If it's a $, it's a variable # and if it's end like it too if elts[0][0] == '$' and elts[0][-1] == '$': macro_name = elts[0][1:-1] self.resource_macros_names.append(macro_name)
def _cut_line(self, line): #punct = '"#$%&\'()*+/<=>?@[\\]^`{|}~' tmp = re.split("[" + string.whitespace + "]+", line, 1) r = [elt for elt in tmp if elt != ''] return r
[docs] def read_config(self, files): # just a first pass to get the cfg_file and all files in a buf res = StringIO() for file in files: # We add a \n (or \r\n) to be sure config files are separated # if the previous does not finish with a line return res.write(os.linesep) res.write('# IMPORTEDFROM=%s' % (file) + os.linesep) if self.read_config_silent == 0: logger.info("[config] opening '%s' configuration file" % file) try: # Open in Universal way for Windows, Mac, Linux fd = open(file, 'rU') buf = fd.readlines() fd.close() self.config_base_dir = os.path.dirname(file) except IOError, exp: logger.error("[config] cannot open config file '%s' for reading: %s" % (file, exp)) # The configuration is invalid because we have a bad file! self.conf_is_correct = False continue for line in buf: line = line.decode('utf8', 'replace') res.write(line) if line.endswith('\n'): line = line[:-1] line = line.strip() if re.search("^cfg_file", line) or re.search("^resource_file", line): elts = line.split('=', 1) if os.path.isabs(elts[1]): cfg_file_name = elts[1] else: cfg_file_name = os.path.join(self.config_base_dir, elts[1]) cfg_file_name = cfg_file_name.strip() try: fd = open(cfg_file_name, 'rU') if self.read_config_silent == 0: logger.info("Processing object config file '%s'" % cfg_file_name) res.write(os.linesep + '# IMPORTEDFROM=%s' % (cfg_file_name) + os.linesep) res.write(fd.read().decode('utf8', 'replace')) # Be sure to add a line return so we won't mix files res.write(os.linesep) fd.close() except IOError, exp: logger.error("Cannot open config file '%s' for reading: %s" % (cfg_file_name, exp)) # The configuration is invalid because we have a bad file! self.conf_is_correct = False elif re.search("^cfg_dir", line): elts = line.split('=', 1) if os.path.isabs(elts[1]): cfg_dir_name = elts[1] else: cfg_dir_name = os.path.join(self.config_base_dir, elts[1]) # Ok, look if it's really a directory if not os.path.isdir(cfg_dir_name): logger.error("Cannot open config dir '%s' for reading" % cfg_dir_name) self.conf_is_correct = False # Look for .pack file into it :) self.packs_dirs.append(cfg_dir_name) # Now walk for it. # BEWARE : we can follow symlinks only for python 2.6 and higher args = {} if sys.version_info >= (2, 6): args['followlinks'] = True for root, dirs, files in os.walk(cfg_dir_name, **args): for file in files: if re.search("\.cfg$", file): if self.read_config_silent == 0: logger.info("Processing object config file '%s'" % os.path.join(root, file)) try: res.write(os.linesep + '# IMPORTEDFROM=%s' % (os.path.join(root, file)) + os.linesep) fd = open(os.path.join(root, file), 'rU') res.write(fd.read().decode('utf8', 'replace')) # Be sure to separate files data res.write(os.linesep) fd.close() except IOError, exp: logger.error("Cannot open config file '%s' for reading: %s" % (os.path.join(root, file), exp)) # The configuration is invalid # because we have a bad file! self.conf_is_correct = False elif re.search("^triggers_dir", line): elts = line.split('=', 1) if os.path.isabs(elts[1]): trig_dir_name = elts[1] else: trig_dir_name = os.path.join(self.config_base_dir, elts[1]) # Ok, look if it's really a directory if not os.path.isdir(trig_dir_name): logger.error("Cannot open triggers dir '%s' for reading" % trig_dir_name) self.conf_is_correct = False continue # Ok it's a valid one, I keep it self.triggers_dirs.append(trig_dir_name) config = res.getvalue() res.close() return config # self.read_config_buf(res)
[docs] def read_config_buf(self, buf): params = [] objectscfg = {} types = self.__class__.configuration_types for t in types: objectscfg[t] = [] tmp = [] tmp_type = 'void' in_define = False almost_in_define=False continuation_line = False tmp_line = '' lines = buf.split('\n') line_nb = 0 # Keep the line number for the file path for line in lines: if line.startswith("# IMPORTEDFROM="): filefrom = line.split('=')[1] line_nb = 0 # reset the line number too continue line_nb += 1 # Remove comments line = split_semicolon(line)[0].strip() # A backslash means, there is more to come if re.search("\\\s*$", line) is not None: continuation_line = True line = re.sub("\\\s*$", "", line) line = re.sub("^\s+", " ", line) tmp_line += line continue elif continuation_line: # Now the continuation line is complete line = re.sub("^\s+", "", line) line = tmp_line + line tmp_line = '' continuation_line = False # } alone in a line means stop the object reading if re.search("^\s*}\s*$", line) is not None: in_define = False # { alone in a line can mean start object reading if re.search("^\s*{\s*$", line) is not None and almost_in_define: almost_in_define=False in_define = True continue if re.search("^\s*#|^\s*$|^\s*}", line) is not None: pass # A define must be catch and the type save # The old entry must be save before elif re.search("^define", line) is not None: if re.search(".*{.*$", line) is not None: in_define = True else: almost_in_define=True if tmp_type not in objectscfg: objectscfg[tmp_type] = [] objectscfg[tmp_type].append(tmp) tmp = [] tmp.append("imported_from " + filefrom+':%d'%line_nb) # Get new type elts = re.split('\s', line) # Maybe there was space before and after the type # so we must get all and strip it tmp_type = ' '.join(elts[1:]).strip() tmp_type = tmp_type.split('{')[0].strip() else: if in_define: tmp.append(line) else: params.append(line) # Maybe the type of the last element is unknown, declare it if not tmp_type in objectscfg: objectscfg[tmp_type] = [] objectscfg[tmp_type].append(tmp) objects = {} #print "Params", params self.load_params(params) # And then update our MACRO dict self.fill_resource_macros_names_macros() for type in objectscfg: objects[type] = [] for items in objectscfg[type]: tmp = {} for line in items: elts = self._cut_line(line) if elts == []: continue prop = elts[0] if not prop in tmp: tmp[prop] = [] value = ' '.join(elts[1:]) tmp[prop].append(value) if tmp != {}: objects[type].append(tmp) return objects # We need to have some ghost objects like # the check_command bp_rule for business # correlator rules
[docs] def add_ghost_objects(self, raw_objects): bp_rule = {'command_name': 'bp_rule', 'command_line': 'bp_rule'} raw_objects['command'].append(bp_rule) host_up = {'command_name': '_internal_host_up', 'command_line': '_internal_host_up'} raw_objects['command'].append(host_up) echo_obj = {'command_name': '_echo', 'command_line': '_echo'} raw_objects['command'].append(echo_obj) # We've got raw objects in string, now create real Instances
[docs] def create_objects(self, raw_objects): """ Create real 'object' from dicts of prop/value """ types_creations = self.__class__.types_creations # some types are already created in this time early_created_types = self.__class__.early_created_types # Before really create the objects, we add # ghost ones like the bp_rule for correlation self.add_ghost_objects(raw_objects) for t in types_creations: if t not in early_created_types: self.create_objects_for_type(raw_objects, t)
[docs] def create_objects_for_type(self, raw_objects, type): types_creations = self.__class__.types_creations t = type # Ex: the above code do for timeperiods: # timeperiods = [] # for timeperiodcfg in objects['timeperiod']: # t = Timeperiod(timeperiodcfg) # t.clean() # timeperiods.append(t) # self.timeperiods = Timeperiods(timeperiods) (cls, clss, prop) = types_creations[t] # List where we put objects lst = [] for obj_cfg in raw_objects[t]: # We create the object o = cls(obj_cfg) lst.append(o) # we create the objects Class and we set it in prop setattr(self, prop, clss(lst)) # Here arbiter and modules objects should be prepare and link # before all others types
[docs] def early_arbiter_linking(self): """ Prepare the arbiter for early operations """ # Should look at hacking command_file module first self.hack_old_nagios_parameters_for_arbiter() self.modules.create_reversed_list() if len(self.arbiters) == 0: logger.warning("There is no arbiter, I add one in localhost:7770") a = ArbiterLink({'arbiter_name': 'Default-Arbiter', 'host_name': socket.gethostname(), 'address': 'localhost', 'port': '7770', 'spare': '0'}) self.arbiters = ArbiterLinks([a]) # First fill default self.arbiters.fill_default() self.modules.fill_default() #print "****************** Pythonize ******************" self.arbiters.pythonize() #print "****************** Linkify ******************" self.arbiters.linkify(self.modules) self.modules.linkify() # We will load all triggers .trig files from all triggers_dir
[docs] def load_triggers(self): for p in self.triggers_dirs: self.triggers.load_file(p) # We will load all packs .pack files from all packs_dirs
[docs] def load_packs(self): for p in self.packs_dirs: self.packs.load_file(p) # We use linkify to make the config more efficient: elements will be # linked, like pointers. For example, a host will have it's service, # and contacts directly in it's properties # REMEMBER: linkify AFTER explode...
[docs] def linkify(self): """ Make 'links' between elements, like a host got a services list with all it's services in it """ # First linkify myself like for some global commands self.linkify_one_command_with_commands(self.commands, 'ocsp_command') self.linkify_one_command_with_commands(self.commands, 'ochp_command') self.linkify_one_command_with_commands(self.commands, 'host_perfdata_command') self.linkify_one_command_with_commands(self.commands, 'service_perfdata_command') #print "Hosts" # link hosts with timeperiods and commands self.hosts.linkify(self.timeperiods, self.commands, \ self.contacts, self.realms, \ self.resultmodulations, self.businessimpactmodulations, \ self.escalations, self.hostgroups, self.triggers, self.checkmodulations, self.macromodulations ) self.hostsextinfo.merge(self.hosts) # Do the simplify AFTER explode groups #print "Hostgroups" # link hostgroups with hosts self.hostgroups.linkify(self.hosts, self.realms) #print "Services" # link services with other objects self.services.linkify(self.hosts, self.commands, \ self.timeperiods, self.contacts,\ self.resultmodulations, self.businessimpactmodulations, \ self.escalations, self.servicegroups, self.triggers, self.checkmodulations, self.macromodulations ) self.servicesextinfo.merge(self.services) #print "Service groups" # link servicegroups members with services self.servicegroups.linkify(self.services) # link notificationways with timeperiods and commands self.notificationways.linkify(self.timeperiods, self.commands) # link notificationways with timeperiods and commands self.checkmodulations.linkify(self.timeperiods, self.commands) # Link with timeperiods self.macromodulations.linkify(self.timeperiods) #print "Contactgroups" # link contacgroups with contacts self.contactgroups.linkify(self.contacts) #print "Contacts" # link contacts with timeperiods and commands self.contacts.linkify(self.timeperiods, self.commands, self.notificationways) #print "Timeperiods" # link timeperiods with timeperiods (exclude part) self.timeperiods.linkify() #print "Servicedependency" self.servicedependencies.linkify(self.hosts, self.services, self.timeperiods) #print "Hostdependency" self.hostdependencies.linkify(self.hosts, self.timeperiods) #print "Resultmodulations" self.resultmodulations.linkify(self.timeperiods) self.businessimpactmodulations.linkify(self.timeperiods) #print "Escalations" self.escalations.linkify(self.timeperiods, self.contacts, \ self.services, self.hosts) # Link discovery commands self.discoveryruns.linkify(self.commands) #print "Realms" self.realms.linkify() #print "Schedulers and satellites" # Link all links with realms #self.arbiters.linkify(self.modules) self.schedulers.linkify(self.realms, self.modules) self.brokers.linkify(self.realms, self.modules) self.receivers.linkify(self.realms, self.modules) self.reactionners.linkify(self.realms, self.modules) self.pollers.linkify(self.realms, self.modules) # Ok, now update all realms with backlinks of # satellites self.realms.prepare_for_satellites_conf() # Some elements are maybe set as wrong after a is_correct, so clean them # if possible
[docs] def clean(self): self.services.clean() # In the scheduler we need to relink the commandCall with # the real commands
[docs] def late_linkify(self): props = ['ocsp_command', 'ochp_command', 'service_perfdata_command', 'host_perfdata_command'] for prop in props: cc = getattr(self, prop, None) if cc: cc.late_linkify_with_command(self.commands) # But also other objects like hosts and services self.hosts.late_linkify_h_by_commands(self.commands) self.services.late_linkify_s_by_commands(self.commands) self.contacts.late_linkify_c_by_commands(self.commands) # Some properties are dangerous to be send like that # like realms linked in hosts. Realms are too big to send (too linked) # We are also pre-serializing the confs so the sending phase will # be quicker.
[docs] def prepare_for_sending(self): # Preparing hosts and hostgroups for sending. Some properties # should be "flatten" before sent, like .realm object that should # be changed into names self.hosts.prepare_for_sending() self.hostgroups.prepare_for_sending() t1 = time.time() logger.info('[Arbiter] Serializing the configurations...') # There are two ways of configuration serializing # One if to use the serial way, the other is with use_multiprocesses_serializer # to call to sub-wrokers to do the job. # TODO : enable on windows? I'm not sure it will work, must give a test if os.name == 'nt' or not self.use_multiprocesses_serializer: logger.info('Using the default serialization pass') for r in self.realms: for (i, conf) in r.confs.iteritems(): # Remember to protect the local conf hostgroups too! conf.hostgroups.prepare_for_sending() logger.debug('[%s] Serializing the configuration %d' % (r.get_name(), i)) t0 = time.time() r.serialized_confs[i] = cPickle.dumps(conf, cPickle.HIGHEST_PROTOCOL) logger.debug("[config] time to serialize the conf %s:%s is %s" % (r.get_name(), i, time.time() - t0)) logger.debug("PICKLE LEN : %d" % len(r.serialized_confs[i])) # Now pickle the whole conf, for easy and quick spare send t0 = time.time() whole_conf_pack = cPickle.dumps(self, cPickle.HIGHEST_PROTOCOL) logger.debug("[config] time to serialize the global conf : %s" % (time.time() - t0)) self.whole_conf_pack = whole_conf_pack print "TOTAL serializing in", time.time() - t1 else: logger.info('Using the multiprocessing serialization pass') t1 = time.time() # We ask a manager to manage the communication with our children m = Manager() # The list will got all the strings from the children q = m.list() for r in self.realms: processes = [] for (i, conf) in r.confs.iteritems(): # This function will be called by the children, and will give # us the pickle result def Serialize_config(q, rname, i, conf): # Remember to protect the local conf hostgroups too! conf.hostgroups.prepare_for_sending() logger.debug('[%s] Serializing the configuration %d' % (rname, i)) t0 = time.time() res = cPickle.dumps(conf, cPickle.HIGHEST_PROTOCOL) logger.debug("[config] time to serialize the conf %s:%s is %s" % (rname, i, time.time() - t0)) q.append((i, res)) # Prepare a sub-process that will manage the pickle computation p = Process(target=Serialize_config, name="serializer-%s-%d" %(r.get_name(), i) , args=(q, r.get_name(), i, conf)) p.start() processes.append( (i, p) ) # Here all sub-processes are launched for this realm, now wait for them to finish while len(processes) != 0: to_del = [] for (i, p) in processes: if p.exitcode is not None: to_del.append((i, p)) # remember to join() so the children can die p.join() for (i, p) in to_del: logger.debug("The sub process %s is done with the return code %d" % (p.name, p.exitcode)) processes.remove( (i, p ) ) # Don't be too quick to poll! time.sleep(0.1) # Check if we got the good number of configuration, maybe one of the cildren got problems? if len(q) != len(r.confs): logger.error("Something goes wrong in the configuration serializations, please restart Shinken Arbiter") sys.exit(2) # Now get the serialized configuration and saved them into self for (i, cfg) in q: r.serialized_confs[i] = cfg print "TOTAL TIME", time.time() - t1 # Now pickle the whole configuration into one big pickle object, for the arbiter spares whole_queue = m.list() t0 = time.time() # The function that just compute the whole conf pickle string, but n a children def create_whole_conf_pack(whole_queue, self): logger.debug("[config] sub processing the whole configuration pack creation") whole_queue.append(cPickle.dumps(self, cPickle.HIGHEST_PROTOCOL)) logger.debug("[config] sub processing the whole configuration pack creation finished") # Go for it p = Process(target=create_whole_conf_pack, args=(whole_queue, self), name='serializer-whole-configuration') p.start() # Wait for it to die while p.exitcode is None: time.sleep(0.1) p.join() # Maybe we don't have our result? if len(whole_queue) != 1: logger.error("Something goes wrong in the whole configuration pack creation, please restart Shinken Arbiter") sys.exit(2) #Get it and save it self.whole_conf_pack = whole_queue.pop() logger.debug("[config] time to serialize the global conf : %s" % (time.time() - t0)) print "TOTAL serializing iin", time.time() - t2 # Shutdown the manager, the sub-process should be gone now m.shutdown()
[docs] def dump(self): print "Slots", Service.__slots__ print 'Hosts:' for h in self.hosts: print '\t', h.get_name(), h.contacts print 'Services:' for s in self.services: print '\t', s.get_name(), s.contacts # It's used to change Nagios2 names to Nagios3 ones # For hosts and services
[docs] def old_properties_names_to_new(self): super(Config, self).old_properties_names_to_new() self.hosts.old_properties_names_to_new() self.services.old_properties_names_to_new() self.notificationways.old_properties_names_to_new() self.contacts.old_properties_names_to_new() # It's used to warn about useless parameter and print why it's not use.
[docs] def notice_about_useless_parameters(self): if not self.disable_old_nagios_parameters_whining: properties = self.__class__.properties for prop, entry in properties.items(): if isinstance(entry, UnusedProp): logger.warning("The parameter %s is useless and can be removed from the configuration (Reason: %s)" % (prop, entry.text)) # It's used to raise warning if the user got parameter # that we do not manage from now
[docs] def warn_about_unmanaged_parameters(self): properties = self.__class__.properties unmanaged = [] for prop, entry in properties.items(): if not entry.managed and hasattr(self, prop): if entry.help: s = "%s: %s" % (prop, entry.help) else: s = prop unmanaged.append(s) if len(unmanaged) != 0: mailing_list_uri = "https://lists.sourceforge.net/lists/listinfo/shinken-devel" logger.warning("The following parameter(s) are not currently managed.") for s in unmanaged: logger.info(s) logger.warning("Unmanaged configuration statement, do you really need it? Ask for it on the developer mailinglist %s or submit a pull request on the Shinken github " % mailing_list_uri) # Overrides specific instances properties
[docs] def override_properties(self): self.services.override_properties(self.hosts) # Use to fill groups values on hosts and create new services # (for host group ones)
[docs] def explode(self): # first elements, after groups #print "Contacts" self.contacts.explode(self.contactgroups, self.notificationways) #print "Contactgroups" self.contactgroups.explode() #print "Hosts" self.hosts.explode(self.hostgroups, self.contactgroups, self.triggers) #print "Hostgroups" self.hostgroups.explode() #print "Services" #print "Initially got nb of services: %d" % len(self.services.items) self.services.explode(self.hosts, self.hostgroups, self.contactgroups, self.servicegroups, self.servicedependencies, self.triggers) #print "finally got nb of services: %d" % len(self.services.items) #print "Servicegroups" self.servicegroups.explode() #print "Timeperiods" self.timeperiods.explode() self.hostdependencies.explode(self.hostgroups) #print "Servicedependency" self.servicedependencies.explode(self.hostgroups) # Serviceescalations hostescalations will create new escalations self.serviceescalations.explode(self.escalations) self.hostescalations.explode(self.escalations) self.escalations.explode(self.hosts, self.hostgroups, self.contactgroups) # Now the architecture part #print "Realms" self.realms.explode() # Remove elements will the same name, so twins :) # In fact only services should be acceptable with twins
[docs] def remove_twins(self): #self.hosts.remove_twins() self.services.remove_twins() #self.contacts.remove_twins() #self.timeperiods.remove_twins() # Dependencies are important for scheduling # This function create dependencies linked between elements.
[docs] def apply_dependencies(self): self.hosts.apply_dependencies() self.services.apply_dependencies() # Use to apply inheritance (template and implicit ones) # So elements will have their configured properties
[docs] def apply_inheritance(self): # inheritance properties by template #print "Hosts" self.hosts.apply_inheritance() #print "Contacts" self.contacts.apply_inheritance() #print "Services" self.services.apply_inheritance(self.hosts) #print "Servicedependencies" self.servicedependencies.apply_inheritance(self.hosts) #print "Hostdependencies" self.hostdependencies.apply_inheritance() # Also timeperiods self.timeperiods.apply_inheritance() # Also "Hostextinfo" self.hostsextinfo.apply_inheritance() # Also "Serviceextinfo" self.servicesextinfo.apply_inheritance() # Now escalations too self.serviceescalations.apply_inheritance() self.hostescalations.apply_inheritance() self.escalations.apply_inheritance() # Use to apply implicit inheritance
[docs] def apply_implicit_inheritance(self): #print "Services" self.services.apply_implicit_inheritance(self.hosts) # will fill properties for elements so they will have all theirs properties
[docs] def fill_default(self): # Fill default for config (self) super(Config, self).fill_default() self.hosts.fill_default() self.hostgroups.fill_default() self.contacts.fill_default() self.contactgroups.fill_default() self.notificationways.fill_default() self.checkmodulations.fill_default() self.macromodulations.fill_default() self.services.fill_default() self.servicegroups.fill_default() self.resultmodulations.fill_default() self.businessimpactmodulations.fill_default() self.hostsextinfo.fill_default() self.servicesextinfo.fill_default() # Now escalations self.escalations.fill_default() # Also fill default of host/servicedep objects self.servicedependencies.fill_default() self.hostdependencies.fill_default() # Discovery part self.discoveryrules.fill_default() self.discoveryruns.fill_default() # first we create missing sat, so no other sat will # be created after this point self.fill_default_satellites() # now we have all elements, we can create a default # realm if need and it will be tagged to sat that do # not have an realm self.fill_default_realm() self.realms.fill_default() # also put default inside the realms themselves self.reactionners.fill_default() self.pollers.fill_default() self.brokers.fill_default() self.receivers.fill_default() self.schedulers.fill_default() # The arbiters are already done. # self.arbiters.fill_default() # Now fill some fields we can predict (like address for hosts) self.fill_predictive_missing_parameters() # Here is a special functions to fill some special # properties that are not filled and should be like # address for host (if not set, put host_name)
[docs] def fill_predictive_missing_parameters(self): self.hosts.fill_predictive_missing_parameters() # Will check if a realm is defined, if not # Create a new one (default) and tag everyone that do not have # a realm prop to be put in this realm
[docs] def fill_default_realm(self): if len(self.realms) == 0: # Create a default realm with default value =1 # so all hosts without realm will be link with it default = Realm({'realm_name': 'Default', 'default': '1'}) self.realms = Realms([default]) logger.warning("No realms defined, I add one at %s" % default.get_name()) lists = [self.pollers, self.brokers, self.reactionners, self.receivers, self.schedulers] for l in lists: for elt in l: if not hasattr(elt, 'realm'): elt.realm = 'Default' logger.info("Tagging %s with realm %s" % (elt.get_name(), default.get_name())) # If a satellite is missing, we add them in the localhost # with defaults values
[docs] def fill_default_satellites(self): if len(self.schedulers) == 0: logger.warning("No scheduler defined, I add one at localhost:7768") s = SchedulerLink({'scheduler_name': 'Default-Scheduler', 'address': 'localhost', 'port': '7768'}) self.schedulers = SchedulerLinks([s]) if len(self.pollers) == 0: logger.warning("No poller defined, I add one at localhost:7771") p = PollerLink({'poller_name': 'Default-Poller', 'address': 'localhost', 'port': '7771'}) self.pollers = PollerLinks([p]) if len(self.reactionners) == 0: logger.warning("No reactionner defined, I add one at localhost:7769") r = ReactionnerLink({'reactionner_name': 'Default-Reactionner', 'address': 'localhost', 'port': '7769'}) self.reactionners = ReactionnerLinks([r]) if len(self.brokers) == 0: logger.warning("No broker defined, I add one at localhost:7772") b = BrokerLink({'broker_name': 'Default-Broker', 'address': 'localhost', 'port': '7772', 'manage_arbiters': '1'}) self.brokers = BrokerLinks([b]) # Return if one broker got a module of type: mod_type
[docs] def got_broker_module_type_defined(self, mod_type): for b in self.brokers: for m in b.modules: if hasattr(m, 'module_type') and m.module_type == mod_type: return True return False # return if one scheduler got a module of type: mod_type
[docs] def got_scheduler_module_type_defined(self, mod_type): for b in self.schedulers: for m in b.modules: if hasattr(m, 'module_type') and m.module_type == mod_type: return True return False # return if one arbiter got a module of type: mod_type # but this time it's tricky: the python pass is not done! # so look with strings!
[docs] def got_arbiter_module_type_defined(self, mod_type): for a in self.arbiters: # Do like the linkify will do after.... for m in getattr(a, 'modules', '').split(','): # So look at what the arbiter try to call as module m = m.strip() # Ok, now look in modules... for mod in self.modules: # try to see if this module is the good type if getattr(mod, 'module_type', '').strip() == mod_type.strip(): # if so, the good name? if getattr(mod, 'module_name', '').strip() == m: return True return False # Will ask for each host/service if the # check_command is a bp rule. If so, it will create # a tree structures with the rules
[docs] def create_business_rules(self): self.hosts.create_business_rules(self.hosts, self.services) self.services.create_business_rules(self.hosts, self.services) # Will fill dep list for business rules
[docs] def create_business_rules_dependencies(self): self.hosts.create_business_rules_dependencies() self.services.create_business_rules_dependencies() # It's used to hack some old Nagios parameters like # log_file or status_file: if they are present in # the global configuration and there is no such modules # in a Broker, we create it on the fly for all Brokers
[docs] def hack_old_nagios_parameters(self): """ Create some 'modules' from all nagios parameters if they are set and the modules are not created """ # We list all modules we will add to brokers mod_to_add = [] mod_to_add_to_schedulers = [] # For status_dat if hasattr(self, 'status_file') and self.status_file != '' and hasattr(self, 'object_cache_file'): # Ok, the user put such a value, we must look # if he forget to put a module for Brokers got_status_dat_module = self.got_broker_module_type_defined('status_dat') # We need to create the module on the fly? if not got_status_dat_module: data = {'object_cache_file': self.object_cache_file, 'status_file': self.status_file, 'module_name': 'Status-Dat-Autogenerated', 'module_type': 'status_dat'} mod = Module(data) mod.status_update_interval = getattr(self, 'status_update_interval', 15) mod_to_add.append(mod) # Now the log_file if hasattr(self, 'log_file') and self.log_file != '': # Ok, the user put such a value, we must look # if he forget to put a module for Brokers got_simple_log_module = self.got_broker_module_type_defined('simple_log') # We need to create the module on the fly? if not got_simple_log_module: data = {'module_type': 'simple_log', 'path': self.log_file, 'archive_path': self.log_archive_path, 'module_name': 'Simple-log-Autogenerated'} mod = Module(data) mod_to_add.append(mod) # Now the syslog facility if self.use_syslog: # Ok, the user want a syslog logging, why not after all got_syslog_module = self.got_broker_module_type_defined('syslog') # We need to create the module on the fly? if not got_syslog_module: data = {'module_type': 'syslog', 'module_name': 'Syslog-Autogenerated'} mod = Module(data) mod_to_add.append(mod) # Now the service_perfdata module if self.service_perfdata_file != '': # Ok, we've got a path for a service perfdata file got_service_perfdata_module = self.got_broker_module_type_defined('service_perfdata') # We need to create the module on the fly? if not got_service_perfdata_module: data = {'module_type': 'service_perfdata', 'module_name': 'Service-Perfdata-Autogenerated', 'path': self.service_perfdata_file, 'mode': self.service_perfdata_file_mode, 'template': self.service_perfdata_file_template} mod = Module(data) mod_to_add.append(mod) # Now the old retention file module if self.state_retention_file != '' and self.retention_update_interval != 0: # Ok, we've got a old retention file got_retention_file_module = self.got_scheduler_module_type_defined('nagios_retention_file') # We need to create the module on the fly? if not got_retention_file_module: data = {'module_type': 'nagios_retention_file', 'module_name': 'Nagios-Retention-File-Autogenerated', 'path': self.state_retention_file} mod = Module(data) mod_to_add_to_schedulers.append(mod) # Now the host_perfdata module if self.host_perfdata_file != '': # Ok, we've got a path for a host perfdata file got_host_perfdata_module = self.got_broker_module_type_defined('host_perfdata') # We need to create the module on the fly? if not got_host_perfdata_module: data = {'module_type': 'host_perfdata', 'module_name': 'Host-Perfdata-Autogenerated', 'path': self.host_perfdata_file, 'mode': self.host_perfdata_file_mode, 'template': self.host_perfdata_file_template} mod = Module(data) mod_to_add.append(mod) # We add them to the brokers if we need it if mod_to_add != []: logger.warning("I autogenerated some Broker modules, please look at your configuration") for m in mod_to_add: logger.warning("The module %s is autogenerated" % m.module_name) for b in self.brokers: b.modules.append(m) # Then for schedulers if mod_to_add_to_schedulers != []: logger.warning("I autogenerated some Scheduler modules, please look at your configuration") for m in mod_to_add_to_schedulers: logger.warning("The module %s is autogenerated" % m.module_name) for b in self.schedulers: b.modules.append(m) # It's used to hack some old Nagios parameters like # but for the arbiter, so very early in the run
[docs] def hack_old_nagios_parameters_for_arbiter(self): """ Create some 'modules' from all nagios parameters if they are set and the modules are not created """ # We list all modules we will add to arbiters mod_to_add = [] # For command_file if getattr(self, 'command_file', '') != '': # Ok, the user put such a value, we must look # if he forget to put a module for arbiters got_named_pipe_module = self.got_arbiter_module_type_defined('named_pipe') # We need to create the module on the fly? if not got_named_pipe_module: data = {'command_file': self.command_file, 'module_name': 'NamedPipe-Autogenerated', 'module_type': 'named_pipe'} mod = Module(data) mod_to_add.append((mod, data)) # We add them to the brokers if we need it if mod_to_add != []: logger.warning("I autogenerated some Arbiter modules, please look at your configuration") for (mod, data) in mod_to_add: logger.warning("Module %s was autogenerated" % data['module_name']) for a in self.arbiters: a.modules = ','.join([getattr(a, 'modules', ''), data['module_name']]) self.modules.items[mod.id] = mod # Set our timezone value and give it too to unset satellites
[docs] def propagate_timezone_option(self): if self.use_timezone != '': # first apply myself os.environ['TZ'] = self.use_timezone time.tzset() tab = [self.schedulers, self.pollers, self.brokers, self.receivers, self.reactionners] for t in tab: for s in t: if s.use_timezone == 'NOTSET': setattr(s, 'use_timezone', self.use_timezone) # Link templates with elements
[docs] def linkify_templates(self): """ Like for normal object, we link templates with each others """ self.hosts.linkify_templates() self.contacts.linkify_templates() self.services.linkify_templates() self.servicedependencies.linkify_templates() self.hostdependencies.linkify_templates() self.timeperiods.linkify_templates() self.hostsextinfo.linkify_templates() self.servicesextinfo.linkify_templates() self.escalations.linkify_templates() # But also old srv and host escalations self.serviceescalations.linkify_templates() self.hostescalations.linkify_templates() # Reversed list is a dist with name for quick search by name
[docs] def create_reversed_list(self): """ Create quick search lists for objects """ self.hosts.create_reversed_list() self.hostgroups.create_reversed_list() self.contacts.create_reversed_list() self.contactgroups.create_reversed_list() self.notificationways.create_reversed_list() self.checkmodulations.create_reversed_list() self.macromodulations.create_reversed_list() self.services.create_reversed_list() self.servicegroups.create_reversed_list() self.timeperiods.create_reversed_list() #self.modules.create_reversed_list() self.resultmodulations.create_reversed_list() self.businessimpactmodulations.create_reversed_list() self.escalations.create_reversed_list() self.discoveryrules.create_reversed_list() self.discoveryruns.create_reversed_list() self.commands.create_reversed_list() self.triggers.create_reversed_list() # For services it's a special case # we search for hosts, then for services # it's quicker than search in all services self.services.optimize_service_search(self.hosts) # Some parameters are just not managed like O*HP commands # and regexp capabilities # True: OK # False: error in conf
[docs] def check_error_on_hard_unmanaged_parameters(self): r = True if self.use_regexp_matching: logger.error("use_regexp_matching parameter is not managed.") r &= False #if self.ochp_command != '': # logger.error("ochp_command parameter is not managed.") # r &= False #if self.ocsp_command != '': # logger.error("ocsp_command parameter is not managed.") # r &= False return r # check if elements are correct or not (fill with defaults, etc) # Warning: this function call be called from a Arbiter AND # from and scheduler. The first one got everything, the second # does not have the satellites.
[docs] def is_correct(self): """ Check if all elements got a good configuration """ logger.info('Running pre-flight check on configuration data...') r = self.conf_is_correct # Globally unmanaged parameters if self.read_config_silent == 0: logger.info('Checking global parameters...') if not self.check_error_on_hard_unmanaged_parameters(): r = False logger.error("Check global parameters failed") for x in ('hosts', 'hostgroups', 'contacts', 'contactgroups', 'notificationways', 'escalations', 'services', 'servicegroups', 'timeperiods', 'commands', 'hostsextinfo', 'servicesextinfo', 'checkmodulations', 'macromodulations'): if self.read_config_silent == 0: logger.info('Checking %s...' % (x)) cur = getattr(self, x) if not cur.is_correct(): r = False logger.error("\t%s conf incorrect!!" % (x)) if self.read_config_silent == 0: logger.info('\tChecked %d %s' % (len(cur), x)) # Hosts got a special check for loops if not self.hosts.no_loop_in_parents(): r = False logger.error("Hosts: detected loop in parents ; conf incorrect") for x in ('servicedependencies', 'hostdependencies', 'arbiters', 'schedulers', 'reactionners', 'pollers', 'brokers', 'receivers', 'resultmodulations', 'discoveryrules', 'discoveryruns', 'businessimpactmodulations'): try: cur = getattr(self, x) except: continue if self.read_config_silent == 0: logger.info('Checking %s...' % (x)) if not cur.is_correct(): r = False logger.error("\t%s conf incorrect!!" % (x)) if self.read_config_silent == 0: logger.info('\tChecked %d %s' % (len(cur), x)) # Look that all scheduler got a broker that will take brok. # If there are no, raise an Error for s in self.schedulers: rea = s.realm if rea: if len(rea.potential_brokers) == 0: logger.error("The scheduler %s got no broker in its realm or upper" % s.get_name()) self.add_error("Error: the scheduler %s got no broker in its realm or upper" % s.get_name()) r = False # Check that for each poller_tag of a host, a poller exists with this tag # TODO: need to check that poller are in the good realm too hosts_tag = set() services_tag = set() pollers_tag = set() for h in self.hosts: hosts_tag.add(h.poller_tag) for s in self.services: services_tag.add(s.poller_tag) for p in self.pollers: for t in p.poller_tags: pollers_tag.add(t) if not hosts_tag.issubset(pollers_tag): for tag in hosts_tag.difference(pollers_tag): logger.error("Hosts exist with poller_tag %s but no poller got this tag" % tag) self.add_error("Error: hosts exist with poller_tag %s but no poller got this tag" % tag) r = False if not services_tag.issubset(pollers_tag): for tag in services_tag.difference(pollers_tag): logger.error("Services exist with poller_tag %s but no poller got this tag" % tag) self.add_error("Error: services exist with poller_tag %s but no poller got this tag" % tag) r = False # Check that all hosts involved in business_rules are from the same realm for l in [self.services, self.hosts]: for e in l: if e.got_business_rule: e_ro = e.get_realm() # Something was wrong in the conf, will be raised elsewhere if not e_ro: continue e_r = e_ro.realm_name for elt in e.business_rule.list_all_elements(): r_o = elt.get_realm() # Something was wrong in the conf, will be raised elsewhere if not r_o: continue elt_r = elt.get_realm().realm_name if not elt_r == e_r: logger.error("Business_rule '%s' got hosts from another realm: %s" % (e.get_full_name(), elt_r)) self.add_error("Error: Business_rule '%s' got hosts from another realm: %s" % (e.get_full_name(), elt_r)) r = False self.conf_is_correct = r # We've got strings (like 1) but we want python elements, like True
[docs] def pythonize(self): # call item pythonize for parameters super(Config, self).pythonize() self.hosts.pythonize() self.hostgroups.pythonize() self.hostdependencies.pythonize() self.contactgroups.pythonize() self.contacts.pythonize() self.notificationways.pythonize() self.checkmodulations.pythonize() self.macromodulations.pythonize() self.servicegroups.pythonize() self.services.pythonize() self.servicedependencies.pythonize() self.resultmodulations.pythonize() self.businessimpactmodulations.pythonize() self.escalations.pythonize() self.discoveryrules.pythonize() self.discoveryruns.pythonize() # The arbiters are already done # self.arbiters.pythonize() self.schedulers.pythonize() self.realms.pythonize() self.reactionners.pythonize() self.pollers.pythonize() self.brokers.pythonize() self.receivers.pythonize() # Explode parameters like cached_service_check_horizon in the # Service class in a cached_check_horizon manner, o*hp commands # , etc
[docs] def explode_global_conf(self): clss = [Service, Host, Contact, SchedulerLink, PollerLink, ReactionnerLink, BrokerLink, ReceiverLink, ArbiterLink, HostExtInfo] for cls in clss: cls.load_global_conf(self) # Clean useless elements like templates because they are not needed anymore
[docs] def remove_templates(self): self.hosts.remove_templates() self.contacts.remove_templates() self.services.remove_templates() self.servicedependencies.remove_templates() self.hostdependencies.remove_templates() self.timeperiods.remove_templates() self.discoveryrules.remove_templates() self.discoveryruns.remove_templates() # We will compute simple element md5hash, so we can know # if they changed or not between the restart
[docs] def compute_hash(self): self.hosts.compute_hash() self.contacts.pythonize() self.notificationways.pythonize() self.checkmodulations.pythonize() self.services.pythonize() self.resultmodulations.pythonize() self.businessimpactmodulations.pythonize() self.escalations.pythonize() self.discoveryrules.pythonize() self.discoveryruns.pythonize() # Add an error in the configuration error list so we can print them # all in one place
[docs] def add_error(self, txt): err = txt self.configuration_errors.append(err) self.conf_is_correct = False # Now it's time to show all configuration errors
[docs] def show_errors(self): for err in self.configuration_errors: logger.error(err) # Create packs of hosts and services so in a pack, # all dependencies are resolved # It create a graph. All hosts are connected to their # parents, and hosts without parent are connected to host 'root'. # services are link to the host. Dependencies are managed # REF: doc/pack-creation.png
[docs] def create_packs(self, nb_packs): # We create a graph with host in nodes g = Graph() g.add_nodes(self.hosts) # links will be used for relations between hosts links = set() # Now the relations for h in self.hosts: # Add parent relations for p in h.parents: if p is not None: links.add((p, h)) # Add the others dependencies for (dep, tmp, tmp2, tmp3, tmp4) in h.act_depend_of: links.add((dep, h)) for (dep, tmp, tmp2, tmp3, tmp4) in h.chk_depend_of: links.add((dep, h)) # For services: they are link with their own host but we need # To have the hosts of service dep in the same pack too for s in self.services: for (dep, tmp, tmp2, tmp3, tmp4) in s.act_depend_of: # I don't care about dep host: they are just the host # of the service... if hasattr(dep, 'host'): links.add((dep.host, s.host)) # The other type of dep for (dep, tmp, tmp2, tmp3, tmp4) in s.chk_depend_of: links.add((dep.host, s.host)) # For host/service that are business based, we need to # link them too for s in [s for s in self.services if s.got_business_rule]: for e in s.business_rule.list_all_elements(): if hasattr(e, 'host'): # if it's a service if e.host != s.host: # do not an host with itself links.add((e.host, s.host)) else: # it's already a host if e != s.host: links.add((e, s.host)) # Same for hosts of course for h in [h for h in self.hosts if h.got_business_rule]: for e in h.business_rule.list_all_elements(): if hasattr(e, 'host'): # if it's a service if e.host != h: links.add((e.host, h)) else: # e is a host if e != h: links.add((e, h)) # Now we create links in the graph. With links (set) # We are sure to call the less add_edge for (dep, h) in links: g.add_edge(dep, h) g.add_edge(h, dep) # Access_list from a node il all nodes that are connected # with it: it's a list of ours mini_packs tmp_packs = g.get_accessibility_packs() # Now We find the default realm (must be unique or # BAD THINGS MAY HAPPEN ) default_realm = None for r in self.realms: if hasattr(r, 'default') and r.default: default_realm = r # Now we look if all elements of all packs have the # same realm. If not, not good! for pack in tmp_packs: tmp_realms = set() for elt in pack: if elt.realm is not None: tmp_realms.add(elt.realm) if len(tmp_realms) > 1: self.add_error("Error: the realm configuration of yours hosts is not good because there a more than one realm in one pack (host relations):") for h in pack: if h.realm is None: err = ' the host %s do not have a realm' % h.get_name() self.add_error(err) else: err = ' the host %s is in the realm %s' % (h.get_name(), h.realm.get_name()) self.add_error(err) if len(tmp_realms) == 1: # Ok, good r = tmp_realms.pop() # There is just one element r.packs.append(pack) elif len(tmp_realms) == 0: # Hum.. no realm value? So default Realm if default_realm is not None: default_realm.packs.append(pack) else: err = "Error: some hosts do not have a realm and you do not defined a default realm!" self.add_error(err) for h in pack: err = ' Impacted host: %s ' % h.get_name() self.add_error(err) # The load balancing is for a loop, so all # hosts of a realm (in a pack) will be dispatch # in the schedulers of this realm # REF: doc/pack-agregation.png # Count the numbers of elements in all the realms, to compare it the total number of hosts nb_elements_all_realms = 0 for r in self.realms: #print "Load balancing realm", r.get_name() packs = {} # create roundrobin iterator for id of cfg # So dispatching is loadbalanced in a realm # but add a entry in the roundrobin tourniquet for # every weight point schedulers (so Weight round robin) weight_list = [] no_spare_schedulers = [s for s in r.schedulers if not s.spare] nb_schedulers = len(no_spare_schedulers) # Maybe there is no scheduler in the realm, it's can be a # big problem if there are elements in packs nb_elements = 0 for pack in r.packs: nb_elements += len(pack) nb_elements_all_realms += len(pack) logger.info("Number of hosts in the realm %s: %d " "(distributed in %d linked packs)" % (r.get_name(), nb_elements, len(r.packs))) if nb_schedulers == 0 and nb_elements != 0: err = "The realm %s has hosts but no scheduler!" % r.get_name() self.add_error(err) r.packs = [] # Dumb pack continue packindex = 0 packindices = {} for s in no_spare_schedulers: packindices[s.id] = packindex packindex += 1 for i in xrange(0, s.weight): weight_list.append(s.id) rr = itertools.cycle(weight_list) # We must have nb_schedulers packs for i in xrange(0, nb_schedulers): packs[i] = [] # Try to load the history association dict so we will try to # send the hosts in the same "pack" assoc = {} if os.path.exists(self.pack_distribution_file): logger.debug('Trying to open the distribution file %s' % self.pack_distribution_file) try: f = open(self.pack_distribution_file, 'rb') assoc = cPickle.load(f) f.close() except Exception, exp: logger.warning('Warning: cannot open the distribution file %s: %s' % (self.pack_distribution_file, str(exp))) # Now we explode the numerous packs into nb_packs reals packs: # we 'load balance' them in a roundrobin way for pack in r.packs: valid_value = False old_pack = -1 for elt in pack: #print 'Look for host', elt.get_name(), 'in assoc' old_i = assoc.get(elt.get_name(), -1) #print 'Founded in ASSOC: ', elt.get_name(),old_i # Maybe it's a new, if so, don't count it if old_i == -1: continue # Maybe it is the first we look at, if so, take it's value if old_pack == -1 and old_i != -1: #print 'First value set', elt.get_name(), old_i old_pack = old_i valid_value = True continue if old_i == old_pack: #print 'I found a match between elements', old_i valid_value = True if old_i != old_pack: #print 'Outch found a change sorry', old_i, old_pack valid_value = False #print 'Is valid?', elt.get_name(), valid_value, old_pack i = None # If it's a valid sub pack and the pack id really exist, use it! if valid_value and old_pack in packindices: #print 'Use a old id for pack', old_pack, [h.get_name() for h in pack] i = old_pack else: # take a new one #print 'take a new id for pack', [h.get_name() for h in pack] i = rr.next() for elt in pack: #print 'We got the element', elt.get_full_name(), ' in pack', i, packindices packs[packindices[i]].append(elt) assoc[elt.get_name()] = i try: logger.info('Saving the distribution file %s' % self.pack_distribution_file) f = open(self.pack_distribution_file, 'wb') cPickle.dump(assoc, f) f.close() except Exception, exp: logger.warning('Could not save the distribution file %s: %s' % (self.pack_distribution_file, str(exp))) # Now in packs we have the number of packs [h1, h2, etc] # equal to the number of schedulers. r.packs = packs logger.info("Total number of hosts : %d" % nb_elements_all_realms) if len(self.hosts) != nb_elements_all_realms: logger.warning("There are %d hosts defined, and %d hosts " "dispatched in the realms. Some hosts have " "been ignored" % (len(self.hosts), nb_elements_all_realms)) self.add_error("There are %d hosts defined, and %d hosts " "dispatched in the realms. Some hosts have " "been ignored" % (len(self.hosts), nb_elements_all_realms)) # Use the self.conf and make nb_parts new confs. # nbparts is equal to the number of schedulerlink # New confs are independent with checks. The only communication # That can be need is macro in commands
[docs] def cut_into_parts(self): #print "Scheduler configured:", self.schedulers # I do not care about alive or not. User must have set a spare if need it nb_parts = len([s for s in self.schedulers if not s.spare]) if nb_parts == 0: nb_parts = 1 # We create dummy configurations for schedulers: # they are clone of the master # conf but without hosts and services (because they are dispatched between # theses configurations) self.confs = {} for i in xrange(0, nb_parts): #print "Create Conf:", i, '/', nb_parts -1 cur_conf = self.confs[i] = Config() # Now we copy all properties of conf into the new ones for prop, entry in Config.properties.items(): if entry.managed and not isinstance(entry, UnusedProp): val = getattr(self, prop) setattr(cur_conf, prop, val) #print "Copy", prop, val # we need a deepcopy because each conf # will have new hostgroups cur_conf.id = i cur_conf.commands = self.commands cur_conf.timeperiods = self.timeperiods # Create hostgroups with just the name and same id, but no members new_hostgroups = [] for hg in self.hostgroups: new_hostgroups.append(hg.copy_shell()) cur_conf.hostgroups = Hostgroups(new_hostgroups) cur_conf.notificationways = self.notificationways cur_conf.checkmodulations = self.checkmodulations cur_conf.macromodulations = self.macromodulations cur_conf.contactgroups = self.contactgroups cur_conf.contacts = self.contacts cur_conf.triggers = self.triggers # Create hostgroups with just the name and same id, but no members new_servicegroups = [] for sg in self.servicegroups: new_servicegroups.append(sg.copy_shell()) cur_conf.servicegroups = Servicegroups(new_servicegroups) cur_conf.hosts = [] # will be fill after cur_conf.services = [] # will be fill after # The elements of the others conf will be tag here cur_conf.other_elements = {} # if a scheduler have accepted the conf cur_conf.is_assigned = False logger.info("Creating packs for realms") # Just create packs. There can be numerous ones # In pack we've got hosts and service # packs are in the realms # REF: doc/pack-creation.png self.create_packs(nb_parts) # We've got all big packs and get elements into configurations # REF: doc/pack-agregation.png offset = 0 for r in self.realms: for i in r.packs: pack = r.packs[i] for h in pack: h.pack_id = i self.confs[i+offset].hosts.append(h) for s in h.services: self.confs[i+offset].services.append(s) # Now the conf can be link in the realm r.confs[i+offset] = self.confs[i+offset] offset += len(r.packs) del r.packs # We've nearly have hosts and services. Now we want REALS hosts (Class) # And we want groups too # print "Finishing packs" for i in self.confs: #print "Finishing pack Nb:", i cfg = self.confs[i] # Create ours classes cfg.hosts = Hosts(cfg.hosts) cfg.hosts.create_reversed_list() cfg.services = Services(cfg.services) cfg.services.create_reversed_list() # Fill host groups for ori_hg in self.hostgroups: hg = cfg.hostgroups.find_by_name(ori_hg.get_name()) mbrs = ori_hg.members mbrs_id = [] for h in mbrs: if h is not None: mbrs_id.append(h.id) for h in cfg.hosts: if h.id in mbrs_id: hg.members.append(h) # Fill servicegroup for ori_sg in self.servicegroups: sg = cfg.servicegroups.find_by_name(ori_sg.get_name()) mbrs = ori_sg.members mbrs_id = [] for s in mbrs: if s is not None: mbrs_id.append(s.id) for s in cfg.services: if s.id in mbrs_id: sg.members.append(s) # Now we fill other_elements by host (service are with their host # so they are not tagged) for i in self.confs: for h in self.confs[i].hosts: for j in [j for j in self.confs if j != i]: # So other than i self.confs[i].other_elements[h.get_name()] = i # We tag conf with instance_id for i in self.confs: self.confs[i].instance_id = i random.seed(time.time()) # ...
def lazy(): # let's compute the "USER" properties and macros.. for n in xrange(1, 256): n = str(n) Config.properties['$USER' + str(n) + '$'] = StringProp(default='') Config.macros['USER' + str(n)] = '$USER' + n + '$' lazy() del lazy
Read the Docs v: documentation
Versions
latest
documentation
Downloads
PDF
HTML
Epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.