def hook_save_retention(self, daemon):
"""
main function that is called in the retention creation pass
"""
try:
self.max_workers = cpu_count()
except NotImplementedError:
pass
t0 = time.time()
logger.debug("[MongodbRetention] asking me to update the retention objects")
all_data = daemon.get_retention_data()
processes = []
for i in xrange(self.max_workers):
proc = Process(target=self.job, args=(all_data, i, self.max_workers))
proc.start()
processes.append(proc)
# Allow 30s to join the sub-processes, should be enough
for proc in processes:
proc.join(30)
logger.info("Retention information updated in Mongodb (%.2fs)" % (time.time() - t0))
def inventor(look_at):
# Now really publish it
inventory = CONFIG['paths']['inventory']
logger.debug("dumping inventory %s" % inventory)
# get all sub-direcotries
for d in os.listdir(inventory):
if os.path.exists(os.path.join(inventory, d, 'package.json')):
if not look_at or d in look_at:
print d
# If asked, dump the content.package content
if look_at or d in look_at:
content_p = os.path.join(inventory, d, 'content.json')
if not os.path.exists(content_p):
logger.error('Missing %s file' % content_p)
continue
try:
j = json.loads(open(content_p, 'r').read())
except Exception, exp:
logger.error('Bad %s file "%s"' % (content_p, exp))
continue
for d in j:
s = ''
if d['type'] == '5': # tar direcotry
s += '(d)'
else:
s += '(f)'
s += d['name']
print s
def manage_service_check_result_brok(self, b):
data = b.data
name = "%s.%s" % (
self.illegal_char.sub('_', data['host_name']),
self.illegal_char.sub('_', data['service_description'])
)
post_data = []
post_data.extend(
self.get_check_result_perfdata_points(
b.data['perf_data'],
b.data['last_chk'],
name
)
)
post_data.extend(
self.get_state_update_points(b.data, name)
)
try:
logger.debug("[influxdb broker] Launching: %s" % str(post_data))
except UnicodeEncodeError:
pass
self.buffer.extend(post_data)
def get_page():
commands_list = []
try:
# Getting lists of informations for the commands
time_stamp_list = []
host_name_list = []
service_description_list = []
return_code_list = []
output_list = []
time_stamp_list = request.forms.getall(key="time_stamp")
logger.debug("[WS_Arbiter] time_stamp_list: %s" % (time_stamp_list))
host_name_list = request.forms.getall(key="host_name")
logger.debug("[WS_Arbiter] host_name_list: %s" % (host_name_list))
service_description_list = request.forms.getall(key="service_description")
logger.debug("[WS_Arbiter] service_description_list: %s" % (service_description_list))
return_code_list = request.forms.getall(key="return_code")
logger.debug("[WS_Arbiter] return_code_list: %s" % (return_code_list))
output_list = request.forms.getall(key="output")
logger.debug("[WS_Arbiter] output_list: %s" % (output_list))
commands_list = get_commands(
time_stamp_list, host_name_list, service_description_list, return_code_list, output_list
)
except Exception, e:
logger.error("[WS_Arbiter] failed to get the lists: %s" % str(e))
commands_list = []
def get_live_data_log(self):
"""Like get_live_data, but for log objects"""
# finalize the filter stacks
self.mongo_time_filter_stack.and_elements(self.mongo_time_filter_stack.qsize())
self.mongo_filter_stack.and_elements(self.mongo_filter_stack.qsize())
if self.use_aggressive_sql:
# Be aggressive, get preselected data from sqlite and do less
# filtering in python. But: only a subset of Filter:-attributes
# can be mapped to columns in the logs-table, for the others
# we must use "always-true"-clauses. This can result in
# funny and potentially ineffective sql-statements
mongo_filter_func = self.mongo_filter_stack.get_stack()
else:
# Be conservative, get everything from the database between
# two dates and apply the Filter:-clauses in python
mongo_filter_func = self.mongo_time_filter_stack.get_stack()
dbresult = []
mongo_filter = mongo_filter_func()
logger.debug("[Logstore MongoDB] Mongo filter is %s" % str(mongo_filter))
# We can apply the filterstack here as well. we have columns and filtercolumns.
# the only additional step is to enrich log lines with host/service-attributes
# A timerange can be useful for a faster preselection of lines
filter_element = eval('{ ' + mongo_filter + ' }')
logger.debug("[LogstoreMongoDB] Mongo filter is %s" % str(filter_element))
columns = ['logobject', 'attempt', 'logclass', 'command_name', 'comment', 'contact_name', 'host_name', 'lineno', 'message', 'plugin_output', 'service_description', 'state', 'state_type', 'time', 'type']
if not self.is_connected == CONNECTED:
logger.warning("[LogStoreMongoDB] sorry, not connected")
else:
dbresult = [Logline([(c,) for c in columns], [x[col] for col in columns]) for x in self.db[self.collection].find(filter_element).sort([(u'time', pymongo.ASCENDING), (u'lineno', pymongo.ASCENDING)])]
return dbresult
def add(self, elt):
cls_type = elt.__class__.my_type
if cls_type == 'brok':
# For brok, we TAG brok with our instance_id
elt.instance_id = 0
self.broks_internal_raised.append(elt)
return
elif cls_type == 'externalcommand':
logger.debug("Enqueuing an external command '%s'" % str(ExternalCommand.__dict__))
self.external_commands.append(elt)
# Maybe we got a Message from the modules, it's way to ask something
# like from now a full data from a scheduler for example.
elif cls_type == 'message':
# We got a message, great!
logger.debug(str(elt.__dict__))
if elt.get_type() == 'NeedData':
data = elt.get_data()
# Full instance id means: I got no data for this scheduler
# so give me all dumbass!
if 'full_instance_id' in data:
c_id = data['full_instance_id']
source = elt.source
logger.info('The module %s is asking me to get all initial data from the scheduler %d' % (source, c_id))
# so we just reset the connection and the running_id, it will just get all new things
try:
self.schedulers[c_id]['con'] = None
self.schedulers[c_id]['running_id'] = 0
except KeyError: # maybe this instance was not known, forget it
logger.warning("the module %s ask me a full_instance_id for an unknown ID (%d)!" % (source, c_id))
# Maybe a module tells me that it's dead, I must log it's last words...
if elt.get_type() == 'ICrash':
data = elt.get_data()
logger.error('the module %s just crash! Please look at the traceback:' % data['name'])
logger.error(data['trace'])
def set_value(obj_ref, output=None, perfdata=None, return_code=None):
obj = get_object(obj_ref)
if not obj:
return
output = output or obj.output
perfdata = perfdata or obj.perf_data
if return_code is None:
return_code = obj.state_id
logger.debug("[trigger] Setting %s %s %s for object %s" % (output, perfdata, return_code, obj.get_full_name()))
if perfdata:
output = output + ' | ' + perfdata
now = time.time()
cls = obj.__class__
i = obj.launch_check(now, force=True)
for chk in obj.checks_in_progress:
if chk.id == i:
logger.debug("[trigger] I found the check I want to change")
c = chk
# Now we 'transform the check into a result'
# So exit_status, output and status is eaten by the host
c.exit_status = return_code
c.get_outputs(output, obj.max_plugins_output_length)
c.status = 'waitconsume'
c.check_time = now
# IMPORTANT: tag this check as from a trigger, so we will not
# loop in an infinite way for triggers checks!
c.from_trigger = True
def get_metric_and_value(self, perf_data):
res = []
metrics = PerfDatas(perf_data)
for e in metrics:
try:
logger.debug("[Graphite UI] groking: %s" % str(e))
except UnicodeEncodeError:
pass
name = self.illegal_char.sub('_', e.name)
name = self.multival.sub(r'.*', name)
# get metric value and its thresholds values if they exist
name_value = {name: (e.value, e.uom)}
if e.warning and e.critical:
name_value[name + '_warn'] = e.warning
name_value[name + '_crit'] = e.critical
# bailout if need
if name_value[name] == '':
continue
try:
logger.debug("[Graphite UI] Got in the end: %s, %s" % (name, e.value))
except UnicodeEncodeError:
pass
for key, value in name_value.items():
res.append((key, value))
return res
def manage_log_brok(self, brok):
"""
Parse a Shinken log brok to enqueue a log line for Index insertion
"""
d = date.today()
index_name = self.index_prefix + "-" + d.strftime("%Y.%m.%d")
line = brok.data["log"]
if re.match("^\[[0-9]*\] [A-Z][a-z]*.:", line):
# Match log which NOT have to be stored
logger.warning("[elastic-logs] do not store: %s", line)
return
logline = Logline(line=line)
logline_dict = logline.as_dict()
logline_dict.update({"@timestamp": datetime.utcfromtimestamp(int(logline_dict["time"])).isoformat() + "Z"})
values = {"_index": index_name, "_type": "shinken-logs", "_source": logline_dict}
# values = logline.as_dict()
if logline.logclass != LOGCLASS_INVALID:
logger.debug("[elastic-logs] store log line values: %s", values)
self.logs_cache.append(values)
else:
logger.info("[elastic-logs] This line is invalid: %s", line)
return
def get_new_broks(self, type='scheduler'):
# Get the good links tab for looping..
links = self.get_links_from_type(type)
if links is None:
logger.debug('Type unknown for connection! %s', type)
return
# We check for new check in each schedulers and put
# the result in new_checks
for sched_id in links:
try:
con = links[sched_id]['con']
if con is not None: # None = not initialized
t0 = time.time()
# Before ask a call that can be long, do a simple ping to be sure it is alive
con.get('ping')
tmp_broks = con.get('get_broks', {'bname':self.name}, wait='long')
try:
_t = base64.b64decode(tmp_broks)
_t = zlib.decompress(_t)
tmp_broks = cPickle.loads(_t)
except (TypeError, zlib.error, cPickle.PickleError), exp:
logger.error('Cannot load broks data from %s : %s', links[sched_id]['name'], exp)
links[sched_id]['con'] = None
continue
logger.debug("%s Broks get in %s", len(tmp_broks), time.time() - t0)
for b in tmp_broks.values():
b.instance_id = links[sched_id]['instance_id']
# Ok, we can add theses broks to our queues
self.add_broks_to_queue(tmp_broks.values())
else: # no con? make the connection
self.pynag_con_init(sched_id, type=type)
def show_minemap():
user = app.request.environ['USER']
# Apply search filter if exists ...
search = app.request.query.get('search', "type:host")
if not "type:host" in search:
search = "type:host "+search
logger.debug("[WebUI-worldmap] search parameters '%s'", search)
items = app.datamgr.search_hosts_and_services(search, user, get_impacts=False)
# Fetch elements per page preference for user, default is 25
elts_per_page = app.prefs_module.get_ui_user_preference(user, 'elts_per_page', 25)
# We want to limit the number of elements
step = int(app.request.GET.get('step', elts_per_page))
start = int(app.request.GET.get('start', '0'))
end = int(app.request.GET.get('end', start + step))
# If we overflow, came back as normal
total = len(items)
if start > total:
start = 0
end = step
navi = app.helper.get_navi(total, start, step=step)
return {'navi': navi, 'search_string': search, 'items': items[start:end], 'page': "minemap"}
def open(self):
"""
Connect to the Mongo DB with configured URI.
Execute a command to check if connected on master to activate immediate connection to
the DB because we need to know if DB server is available.
Update log rotation time to force a log rotation
"""
self.con = MongoClient(self.uri, connect=False)
logger.info("[mongo-logs] trying to connect MongoDB: %s", self.uri)
try:
result = self.con.admin.command("ismaster")
logger.info("[mongo-logs] connected to MongoDB, admin: %s", result)
logger.debug("[mongo-logs] server information: %s", self.con.server_info())
self.db = getattr(self.con, self.database)
logger.info("[mongo-logs] connected to the database: %s (%s)", self.database, self.db)
self.is_connected = CONNECTED
self.next_logs_rotation = time.time()
logger.info('[mongo-logs] database connection established')
except ConnectionFailure as e:
logger.error("[mongo-logs] Server is not available: %s", str(e))
return False
except Exception as e:
logger.error("[mongo-logs] Could not open the database", str(e))
raise MongoLogsError
return True
def do_pynag_con_init(self, id, type='scheduler'):
# Get the good links tab for looping..
links = self.get_links_from_type(type)
if links is None:
logger.debug('Type unknown for connection! %s', type)
return
if type == 'scheduler':
# If sched is not active, I do not try to init
# it is just useless
is_active = links[id]['active']
if not is_active:
return
# If we try to connect too much, we slow down our tests
if self.is_connection_try_too_close(links[id]):
return
# Ok, we can now update it
links[id]['last_connection'] = time.time()
# DBG: print "Init connection with", links[id]['uri']
running_id = links[id]['running_id']
# DBG: print "Running id before connection", running_id
uri = links[id]['uri']
try:
con = links[id]['con'] = HTTPClient(uri=uri, strong_ssl=links[id]['hard_ssl_name_check'])
except HTTPExceptions, exp:
# But the multiprocessing module is not compatible with it!
# so we must disable it immediately after
logger.info("Connection problem to the %s %s: %s", type, links[id]['name'], str(exp))
links[id]['con'] = None
return
def post(self, path, args, wait='short'):
size = 0
# Take args, pickle them and then compress the result
for (k,v) in args.iteritems():
args[k] = zlib.compress(cPickle.dumps(v), 2)
size += len(args[k])
# Ok go for it!
logger.debug('Posting to %s: %sB' % (self.uri+path, size))
c = self.con
c.setopt(pycurl.HTTPGET, 0)
c.setopt(c.POST, 1)
# For the TIMEOUT, it will depends if we are waiting for a long query or not
# long:data_timeout, like for huge broks receptions
# short:timeout, like for just "ok" connection
if wait == 'short':
c.setopt(c.TIMEOUT, self.timeout)
else:
c.setopt(c.TIMEOUT, self.data_timeout)
#if proxy:
# c.setopt(c.PROXY, proxy)
# Pycurl want a list of tuple as args
postargs = [(k,v) for (k,v) in args.iteritems()]
c.setopt(c.HTTPPOST, postargs)
c.setopt(c.URL, str(self.uri+path))
# Ok now manage the response
response = StringIO()
c.setopt(pycurl.WRITEFUNCTION, response.write)
#c.setopt(c.VERBOSE, 1)
try:
c.perform()
except pycurl.error, error:
errno, errstr = error
raise HTTPException ('Connexion error to %s : %s' % (self.uri, errstr))
def get_graph_uris(self, elt, graphstart=None, graphend=None, duration=None, source='detail'):
''' Aggregate the get_graph_uris of all the submodules.
The source parameter defines the source of the calling:
Are we displaying graphs for the element detail page (detail),
or a widget in the dashboard (dashboard) ?
If duration is not None, we consider it as a number of seconds to graph and
we call the module get_relative_graphs_uri
If get_relative_graphs_uri is not a module function we compute graphstart and
graphend and we call we call the module get_graphs_uri
If graphstart and graphend are not None, we call the module get_graphs_uri
'''
uris = []
for mod in self.modules:
if not duration:
uris.extend(mod.get_graph_uris(elt, graphstart, graphend, source))
else:
f = getattr(mod, 'get_relative_graph_uris', None)
if f and callable(f):
uris.extend(f(elt, duration, source))
else:
graphend = time.time()
graphstart = graphend - duration
uris.extend(mod.get_graph_uris(elt, graphstart, graphend, source))
logger.debug("[WebUI] Got graphs: %s", uris)
for uri in uris:
uri['img_src'] = '/graph?url=' + urllib.quote(uri['img_src'])
return uris
def allperfs(obj_ref):
""" Get all perfdatas from a service or a host
"""
obj = get_object(obj_ref)
p = PerfDatas(obj.perf_data)
logger.debug("[trigger] I get all perfdatas")
return dict([(metric.name, p[metric.name]) for metric in p])
def linkify_hg_by_realms(self, realms):
# Now we explode the realm value if we've got one
# The group realm must not override a host one (warning?)
for hg in self:
if not hasattr(hg, 'realm'):
continue
# Maybe the value is void?
if not hg.realm.strip():
continue
r = realms.find_by_name(hg.realm.strip())
if r is not None:
hg.realm = r
logger.debug("[hostgroups] %s is in %s realm", hg.get_name(), r.get_name())
else:
err = "the hostgroup %s got an unknown realm '%s'" % (hg.get_name(), hg.realm)
hg.configuration_errors.append(err)
hg.realm = None
continue
for h in hg:
if h is None:
continue
if h.realm is None or h.got_default_realm: # default value not hasattr(h, 'realm'):
logger.debug("[hostgroups] apply a realm %s to host %s from a hostgroup rule (%s)", \
hg.realm.get_name(), h.get_name(), hg.get_name())
h.realm = hg.realm
else:
if h.realm != hg.realm:
logger.warning("[hostgroups] host %s it not in the same realm than it's hostgroup %s", \
h.get_name(), hg.get_name())
def hook_save_retention(self, daemon):
"""
main function that is called in the retention creation pass
"""
logger.debug("[MemcacheRetention] asking me to update the retention objects")
all_data = daemon.get_retention_data()
hosts = all_data['hosts']
services = all_data['services']
# Now the flat file method
for h_name in hosts:
try:
h = hosts[h_name]
key = self.normalize_key("HOST-%s" % h_name)
val = cPickle.dumps(h)
self.mc.set(key, val)
except:
logger.error("[MemcacheRetention] error while saving host %s" % key)
for (h_name, s_desc) in services:
try:
key = self.normalize_key("SERVICE-%s,%s" % (h_name, s_desc))
s = services[(h_name, s_desc)]
val = cPickle.dumps(s)
self.mc.set(key, val)
except:
logger.error("[MemcacheRetention] error while saving service %s" % key)
self.mc.disconnect_all()
logger.info("Retention information updated in Memcache")
def commit_and_rotate_log_db(self):
"""Submit a commit or rotate the complete database file.
This function is called whenever the mainloop doesn't handle a request.
The database updates are committed every second.
Every day at 00:05 the database contents with a timestamp of past days
are moved to their own datafiles (one for each day). We wait until 00:05
because in a distributed environment even after 00:00 (on the broker host)
we might receive data from other hosts with a timestamp dating from yesterday.
"""
if self.read_only:
return
now = time.time()
if self.next_log_db_commit <= now:
self.commit()
logger.debug("[Logstore SQLite] commit.....")
self.next_log_db_commit = now + 1
if self.next_log_db_rotate <= now:
logger.info("[Logstore SQLite] at %s we rotate the database file" % time.asctime(time.localtime(now)))
# Take the current database file
# Move the messages into daily files
self.log_db_do_archive()
today = datetime.date.today()
today0005 = datetime.datetime(today.year, today.month, today.day, 0, 5, 0)
if now < time.mktime(today0005.timetuple()):
nextrotation = today0005
else:
nextrotation = today0005 + datetime.timedelta(days=1)
# See you tomorrow
self.next_log_db_rotate = time.mktime(nextrotation.timetuple())
logger.info("[Logstore SQLite] next rotation at %s " % time.asctime(time.localtime(self.next_log_db_rotate)))
def hook_save_retention(self, daemon):
"""
main function that is called in the retention creation pass
"""
logger.debug("[RedisRetention] asking me to update retention objects")
all_data = daemon.get_retention_data()
hosts = all_data['hosts']
services = all_data['services']
# Now the flat file method
for h_name in hosts:
h = hosts[h_name]
key = self._get_host_key(h_name)
val = cPickle.dumps(h)
if self.expire_time:
self.rc.set(key, val, ex=self.expire_time)
else:
self.rc.set(key, val)
for (h_name, s_desc) in services:
s = services[(h_name, s_desc)]
key = self._get_service_key(h_name, s_desc)
val = cPickle.dumps(s)
if self.expire_time:
self.rc.set(key, val, ex=self.expire_time)
else:
self.rc.set(key, val)
logger.info("Retention information updated in Redis")
def init(self):
"""
Called by Scheduler to say 'let's prepare yourself guy'
"""
logger.debug("Initialization of the memcache module")
#self.return_queue = self.properties['from_queue']
self.mc = memcache.Client(['%s:%s' % (self.server, self.port)], debug=0)
def grab_package(pname):
cprint('Grabbing : ' , end='')
cprint('%s' % pname, 'green')
# Now really publish it
proxy = CONFIG['shinken.io']['proxy']
api_key = CONFIG['shinken.io']['api_key']
# Ok we will push the file with a 10s timeout
c = pycurl.Curl()
c.setopt(c.POST, 0)
c.setopt(c.CONNECTTIMEOUT, 10)
c.setopt(c.TIMEOUT, 10)
if proxy:
c.setopt(c.PROXY, proxy)
c.setopt(c.URL, str('shinken.io/grab/%s' % pname))
response = StringIO()
c.setopt(pycurl.WRITEFUNCTION, response.write)
#c.setopt(c.VERBOSE, 1)
c.perform()
r = c.getinfo(pycurl.HTTP_CODE)
c.close()
if r != 200:
logger.error("There was a critical error : %s" % response.getvalue())
sys.exit(2)
else:
ret = response.getvalue()
logger.debug("CURL result len : %d " % len(ret))
return ret
def get_instance(plugin):
logger.debug("[MySQLImport]: Get MySQL importer instance for plugin %s" % plugin.get_name())
if not MySQLdb:
raise Exception('Missing module python-mysqldb. Please install it.')
host = plugin.host
login = plugin.login
password = plugin.password
database = plugin.database
reqlist = {}
reqlist['hosts'] = getattr(plugin, 'reqhosts', None)
reqlist['commands'] = getattr(plugin, 'reqcommands', None)
reqlist['timeperiods'] = getattr(plugin, 'reqtimeperiods', None)
reqlist['notificationways'] = getattr(plugin, 'reqnotificationways', None)
reqlist['services'] = getattr(plugin, 'reqservices', None)
reqlist['servicegroups'] = getattr(plugin, 'reqservicegroups', None)
reqlist['contacts'] = getattr(plugin, 'reqcontacts', None)
reqlist['contactgroups'] = getattr(plugin, 'reqcontactgroups', None)
reqlist['hostgroups'] = getattr(plugin, 'reqhostgroups', None)
reqlist['hostdependencies'] = getattr(plugin, 'reqhostdependencies', None)
reqlist['servicedependencies'] = getattr(plugin, 'reqservicedependencies', None)
reqlist['realms'] = getattr(plugin, 'reqrealms', None)
reqlist['schedulers'] = getattr(plugin, 'reqschedulers', None)
reqlist['pollers'] = getattr(plugin, 'reqpollers', None)
reqlist['brokers'] = getattr(plugin, 'reqbrokers', None)
reqlist['reactionners'] = getattr(plugin, 'reqreactionners', None)
reqlist['receivers'] = getattr(plugin, 'reqreceivers', None)
instance = MySQL_importer_arbiter(plugin, host, login, password, database, reqlist)
return instance
def process_check_result(self, databuffer, IV):
# 208 is the size of fixed received data ... NSCA packets are 208+512 (720) or 208+4096 (4304)
if not databuffer:
logger.warning("[NSCA] Received an empty NSCA packet")
return
logger.debug("[NSCA] Received NSCA packet: %s", binascii.hexlify(databuffer))
payload_length = len(databuffer) - 208
if payload_length != 512 and payload_length != 4096:
logger.warning("[NSCA] Received packet with unusual payload length: %d.", payload_length)
if self.payload_length != -1 and payload_length != self.payload_length:
logger.warning("[NSCA] Dropping packet with incorrect payload length.")
return
(timestamp, rc, hostname, service, output) = self.read_check_result(databuffer, IV, payload_length)
current_time = time.time()
check_result_age = current_time - timestamp
if timestamp > current_time and self.check_future_packet:
logger.warning("[NSCA] Dropping packet with future timestamp.")
elif check_result_age > self.max_packet_age:
logger.info(
"[NSCA] Dropping packet with stale timestamp - packet was %s seconds old. Timestamp: %s for %s/%s" % \
(check_result_age, timestamp, hostname, service))
else:
self.post_command(timestamp, rc, hostname, service, output)
def get_commands(time_stamps, hosts, services, return_codes, outputs):
"""Composing a command list based on the information received in
POST request"""
commands = []
current_time_stamp = int(time.time())
def _compose_command(t, h, s, r, o):
"""Simple function to create a command from the inputs"""
cmd = ""
if not s or s == "":
cmd = '[%s] PROCESS_HOST_CHECK_RESULT;%s;%s;%s' % (t if t is not None else current_time_stamp, h, r, o)
else:
cmd = '[%s] PROCESS_SERVICE_CHECK_RESULT;%s;%s;%s;%s' % (t if t is not None else current_time_stamp, h, s, r, o)
logger.debug("[Ws_arbiter] CMD: %s" % (cmd))
commands.append(cmd)
# Trivial case: empty commmand list
if (return_codes is None or len(return_codes) == 0):
return commands
# Sanity check: if we get N return codes, we must have N hosts.
# The other values could be None
if (len(return_codes) != len(hosts)):
logger.error("[Ws_arbiter] number of return codes (%d) does not match number of hosts (%d)" % (len(return_codes), len(hosts)))
abort(400, "number of return codes does not match number of hosts")
map(_compose_command, time_stamps, hosts, services, return_codes, outputs)
logger.debug("[Ws_arbiter] commands = %s" % (str(commands)))
return commands
def set_ui_user_preference(self, user, key, value):
if not self.is_connected:
if not self.open():
logger.error("[WebUI-MongoDBPreferences] error during initialization, no database connection!")
return None
if not user:
logger.warning("[WebUI-MongoDBPreferences] error set_ui_user_preference, no user!")
return None
try:
# check a collection exist for this user
u = self.db.ui_user_preferences.find_one({'_id': user.get_name()})
if not u:
# no collection for this user? create a new one
self.db.ui_user_preferences.save({'_id': user.get_name(), key: value})
r = self.db.ui_user_preferences.update({'_id': user.get_name()}, {'$set': {key: value}})
# Maybe there was no doc there, if so, create an empty one
if not r:
# Maybe the user exist, if so, get the whole user entry
u = self.db.ui_user_preferences.find_one({'_id': user.get_name()})
if not u:
logger.debug ("[WebUI-MongoDBPreferences] No user entry for %s, I create a new one", user.get_name())
self.db.ui_user_preferences.save({'_id': user.get_name(), key: value})
else: # ok, it was just the key that was missing, just update it and save it
u[key] = value
logger.debug ("[WebUI-MongoDBPreferences] Just saving the new key in the user pref")
self.db.ui_user_preferences.save(u)
except Exception, e:
logger.warning("[WebUI-MongoDBPreferences] Exception: %s", str(e))
self.is_connected = False
return None
def hook_tick(self, brok):
"""Each second the broker calls the hook_tick function
Every tick try to flush the buffer
"""
if self.buffer == []:
return
# Todo : why we need this?
if self.ticks >= self.tick_limit:
# If the number of ticks where data was not
# sent successfully to the raw socket reaches the buffer limit.
# Reset the buffer and reset the ticks
self.buffer = []
self.ticks = 0
return
# Real memory size
if sum(x.__sizeof__() for x in self.buffer) > self.max_buffer_size:
logger.debug("[RawSocket broker] Buffer size exceeded. I delete %d lines"
% self.lines_deleted)
self.buffer = self.buffer[self.lines_deleted:]
self.ticks += 1
try:
self.con.sendall('\n'.join(self.buffer).encode('UTF-8') + '\n')
except IOError, err:
logger.error("[RawSocket broker] Failed sending to the Raw network socket! IOError:%s"
% str(err))
self.init()
return
def manage_unknown_service_check_result_brok(self, b):
data = b.data
tags = {
"host_name": data['host_name'],
"service_description": data['service_description']
}
post_data = []
post_data.extend(
self.get_check_result_perfdata_points(
b.data['perf_data'],
b.data['time_stamp'],
tags=tags
)
)
try:
logger.debug(
"[influxdb broker] Generated points: %s" % str(post_data))
except UnicodeEncodeError:
pass
self.extend_buffer(post_data)
def do_recheck():
# Getting lists of informations for the commands
time_stamp = request.forms.get('time_stamp', int(time.time()))
host_name = request.forms.get('host_name', '')
service_description = request.forms.get('service_description', '')
logger.debug("[WS_Arbiter] Timestamp '%s' - host: '%s', service: '%s'" % (time_stamp,
host_name,
service_description
)
)
if not host_name:
abort(400, 'Missing parameter host_name')
if service_description:
# SCHEDULE_FORCED_SVC_CHECK;<host_name>;<service_description>;<check_time>
command = '[%s] SCHEDULE_FORCED_SVC_CHECK;%s;%s;%s\n' % (time_stamp,
host_name,
service_description,
time_stamp)
else:
# SCHEDULE_FORCED_HOST_CHECK;<host_name>;<check_time>
command = '[%s] SCHEDULE_FORCED_HOST_CHECK;%s;%s\n' % (time_stamp,
host_name,
time_stamp)
# We check for auth if it's not anonymously allowed
check_auth()
# Adding commands to the main queue()
logger.debug("[WS_Arbiter] command = %s" % command)
ext = ExternalCommand(command)
app.from_q.put(ext)
def manage_initial_broks_done_brok(self, b):
if self.con is None:
return
logger.info("[Active Directory UI] AD/LDAP: manage_initial_broks_done_brok, go for pictures")
searchScope = ldap.SCOPE_SUBTREE
## retrieve all attributes - again adjust to your needs - see documentation for more options
#retrieveAttributes = ["userPrincipalName", "thumbnailPhoto", "samaccountname", "email"]
logger.info("[Active Directory UI] Contacts? %d" % len(self.app.datamgr.get_contacts()))
for c in self.app.datamgr.get_contacts():
logger.debug("[Active Directory UI] Doing photo lookup for contact: %s" % c.get_name())
elts = self.find_contact_entry(c)
if elts is None:
logger.warning("[Active Directory UI] No ldap entry for %s" % c.get_name())
continue
# Ok, try to get photo from the entry
try:
photo = elts[self.photo_attr][0]
try:
p = os.path.join(self.app.photo_dir, c.get_name()+'.jpg')
f = open(p, 'wb')
f.write(photo)
f.close()
logger.info("[Active Directory UI] Photo wrote for %s" % c.get_name())
except Exception, exp:
logger.error("[Active Directory UI] Cannot write %s : %s" % (p, str(exp)))
except KeyError:
logger.warning("[Active Directory UI] No photo for %s" % c.get_name())
def setup_new_conf(self):
conf = self.new_conf
self.new_conf = None
self.cur_conf = conf
# Got our name from the globals
if 'receiver_name' in conf['global']:
name = conf['global']['receiver_name']
else:
name = 'Unnamed receiver'
self.name = name
logger.load_obj(self, name)
self.direct_routing = conf['global']['direct_routing']
g_conf = conf['global']
# If we've got something in the schedulers, we do not want it anymore
for sched_id in conf['schedulers']:
already_got = False
# We can already got this conf id, but with another address
if sched_id in self.schedulers:
new_addr = conf['schedulers'][sched_id]['address']
old_addr = self.schedulers[sched_id]['address']
new_port = conf['schedulers'][sched_id]['port']
old_port = self.schedulers[sched_id]['port']
# Should got all the same to be ok :)
if new_addr == old_addr and new_port == old_port:
already_got = True
if already_got:
logger.info("[%s] We already got the conf %d (%s)" %
(self.name, sched_id,
conf['schedulers'][sched_id]['name']))
wait_homerun = self.schedulers[sched_id]['wait_homerun']
actions = self.schedulers[sched_id]['actions']
external_commands = self.schedulers[sched_id][
'external_commands']
con = self.schedulers[sched_id]['con']
s = conf['schedulers'][sched_id]
self.schedulers[sched_id] = s
if s['name'] in g_conf['satellitemap']:
s.update(g_conf['satellitemap'][s['name']])
proto = 'http'
if s['use_ssl']:
proto = 'https'
uri = '%s://%s:%s/' % (proto, s['address'], s['port'])
self.schedulers[sched_id]['uri'] = uri
if already_got:
self.schedulers[sched_id]['wait_homerun'] = wait_homerun
self.schedulers[sched_id]['actions'] = actions
self.schedulers[sched_id][
'external_commands'] = external_commands
self.schedulers[sched_id]['con'] = con
else:
self.schedulers[sched_id]['wait_homerun'] = {}
self.schedulers[sched_id]['actions'] = {}
self.schedulers[sched_id]['external_commands'] = []
self.schedulers[sched_id]['con'] = None
self.schedulers[sched_id]['running_id'] = 0
self.schedulers[sched_id]['active'] = s['active']
# Do not connect if we are a passive satellite
if self.direct_routing and not already_got:
# And then we connect to it :)
self.pynag_con_init(sched_id)
logger.debug("[%s] Sending us configuration %s" % (self.name, conf))
if not self.have_modules:
self.modules = mods = conf['global']['modules']
self.have_modules = True
logger.info("We received modules %s " % mods)
# Set our giving timezone from arbiter
use_timezone = conf['global']['use_timezone']
if use_timezone != 'NOTSET':
logger.info("Setting our timezone to %s" % use_timezone)
os.environ['TZ'] = use_timezone
time.tzset()
# Now create the external commander. It's just here to dispatch
# the commands to schedulers
e = ExternalCommandManager(None, 'receiver')
e.load_receiver(self)
self.external_command = e
def set_proxy(self, proxy):
if proxy:
logger.debug('PROXY SETTING PROXY %s', proxy)
self.get_con.setopt(pycurl.PROXY, proxy)
self.post_con.setopt(pycurl.PROXY, proxy)
self.put_con.setopt(pycurl.PROXY, proxy)
def check_auth(self, username, password):
""" Check username/password.
If there is submodules, this method calls them one by one until one of them returns
True. If no submodule can authenticate the user, then we try with internal
authentication methods: htpasswd file, then contact password.
This method returns a User object if authentication succeeded, else it returns None
"""
self._user_login = None
self._authenticator = None
self._session = None
self._user_info = None
logger.info("[WebUI] Authenticating user '%s'", username)
self.app.request.environ['MSG'] = "Unable to authenticate a user"
if self.modules:
for mod in self.modules:
try:
logger.info("[WebUI] Authenticating user '%s' with %s",
username, mod.get_name())
if mod.check_auth(username, password):
logger.debug(
"[WebUI] User '%s' is authenticated thanks to %s",
username, mod.get_name())
self._authenticator = mod.get_name()
self._user_login = username
# Session identifier ?
f = getattr(mod, 'get_session', None)
if f and callable(f):
self._session = mod.get_session()
logger.info("[WebUI] User session: %s",
self._session)
# User information ?
f = getattr(mod, 'get_user_info', None)
if f and callable(f):
self._user_info = mod.get_user_info()
logger.info("[WebUI] User info: %s",
self._user_info)
except Exception as exp:
logger.warning("[WebUI] Exception: %s", str(exp))
logger.warning("[WebUI] Back trace: %s",
traceback.format_exc())
if not self._user_login:
logger.info("[WebUI] Internal htpasswd authentication")
if self.app.htpasswd_file and self.check_apache_htpasswd_auth(
username, password):
self._authenticator = 'htpasswd'
self._user_login = username
if not self._user_login:
logger.info("[WebUI] Internal contact authentication")
if self.check_cfg_password_auth(username, password):
self._authenticator = 'contact'
self._user_login = username
if self._user_login:
logger.info("[WebUI] user authenticated thanks to %s",
self._authenticator)
self.app.request.environ['MSG'] = "Welcome to the WebUI"
return self._user_login
return None
def check_apache_htpasswd_auth(self, username, password):
""" Embedded authentication with password in Apache htpasswd file.
Function imported from auth-htpasswd module.
"""
logger.info("[WebUI-auth-htpasswd] Authenticating user '%s'", username)
try:
f = open(self.app.htpasswd_file, 'r')
for line in f.readlines():
line = line.strip()
# Bypass bad lines
if ':' not in line:
continue
if line.startswith('#'):
continue
elts = line.split(':')
name = elts[0]
my_hash = elts[1]
if my_hash[:5] == '$apr1' or my_hash[:3] == '$1$':
h = my_hash.split('$')
magic = h[1]
salt = h[2]
elif my_hash[0] == '$':
h = my_hash.split('$')
magic = h[1]
else:
magic = None
salt = my_hash[:2]
# If we match the username, look at the crypt
if name == username:
if md5_available and magic == 'apr1':
valid_hash = (apache_md5_crypt(password,
salt) == my_hash)
elif md5_available and magic == '1':
valid_hash = (unix_md5_crypt(password,
salt) == my_hash)
elif passlib_available and (magic[0] == '2'):
valid_hash = bcrypt.verify(password, my_hash)
elif passlib_available and magic == '5':
valid_hash = sha256_crypt.verify(password, my_hash)
elif passlib_available and magic == '6':
valid_hash = sha512_crypt.verify(password, my_hash)
elif magic is None:
valid_hash = (crypt.crypt(password, salt) == my_hash)
if valid_hash:
logger.info("[WebUI-auth-htpasswd] Authenticated")
return True
else:
logger.debug(
"[WebUI-auth-htpasswd] Authentication failed, "
"invalid name: %s / %s", name, username)
except Exception as exp:
logger.info(
"[WebUI-auth-htpasswd] Authentication against apache passwd "
"file failed, exception: %s", str(exp))
finally:
try:
f.close()
except Exception:
pass
return False
def user_login():
logger.debug("[WebUI] user login request, remote user enabled: %s: %s",
app.remote_user_enable, app.remote_user_variable)
for header in app.request.headers:
logger.debug("[WebUI] X request header: %s = %s", header,
app.request.headers[header])
err = app.request.GET.get('error', None)
if err:
logger.warning("[WebUI] login page with error message: %s", err)
cookie_value = app.request.get_cookie(app.session_cookie,
secret=app.auth_secret)
if cookie_value:
logger.info("[WebUI] user login request, existing cookie found: %s",
cookie_value)
bottle.redirect(app.get_url("Dashboard"))
elif app.remote_user_enable in ['1', '2']:
logger.debug("[WebUI] user login request, no existing cookie found")
if not err:
user_name = None
if app.remote_user_enable == '1':
logger.debug("[WebUI] search %s in request headers",
app.remote_user_variable)
if app.remote_user_variable in app.request.headers:
user_name = app.request.headers[app.remote_user_variable]
logger.debug(
"[WebUI] remote user found in request headers: %s",
user_name)
if app.remote_user_enable == '2':
logger.debug("[WebUI] search %s in WSGI environment",
app.remote_user_variable)
if app.remote_user_variable in app.request.environ:
user_name = app.request.environ[app.remote_user_variable]
logger.debug(
"[WebUI] remote user found in WSGI environment: %s",
user_name)
if not user_name:
logger.warning(
"[WebUI] remote user is enabled but no authenticated "
"user name was found")
bottle.redirect(app.get_url("GetLogin"))
c = app.datamgr.get_contact(name=user_name)
if c:
cookie_value = {
'login': user_name,
'session': app.user_session,
'info': app.user_info
}
app.response.set_cookie(str(app.session_cookie),
cookie_value,
secret=app.auth_secret,
path='/')
bottle.redirect(app.get_url("Dashboard"))
logger.info("[WebUI] session user message - get: %s",
app.request.environ.get('MSG', 'None...'))
return {
'msg_text': err,
'login_text': app.login_text,
'company_logo': app.company_logo
}
def dispatch(self):
# Ok, we pass at least one time in dispatch, so now errors are True errors
self.first_dispatch_done = True
# If no needed to dispatch, do not dispatch :)
if not self.dispatch_ok:
for r in self.realms:
conf_to_dispatch = [
cfg for cfg in r.confs.values() if not cfg.is_assigned
]
nb_conf = len(conf_to_dispatch)
if nb_conf > 0:
logger.info("Dispatching Realm %s" % r.get_name())
logger.info('[%s] Dispatching %d/%d configurations' %
(r.get_name(), nb_conf, len(r.confs)))
# Now we get in scheds all scheduler of this realm and upper so
# we will send them conf (in this order)
scheds = self.get_scheduler_ordered_list(r)
if nb_conf > 0:
print_string = '[%s] Schedulers order: %s' % (
r.get_name(), ','.join([s.get_name() for s in scheds]))
logger.info(print_string)
# Try to send only for alive members
scheds = [s for s in scheds if s.alive]
# Now we do the real job
# every_one_need_conf = False
for conf in conf_to_dispatch:
logger.info('[%s] Dispatching configuration %s' %
(r.get_name(), conf.id))
# If there is no alive schedulers, not good...
if len(scheds) == 0:
logger.info(
'[%s] but there a no alive schedulers in this realm!'
% r.get_name())
# we need to loop until the conf is assigned
# or when there are no more schedulers available
while True:
try:
sched = scheds.pop()
except IndexError: # No more schedulers.. not good, no loop
# need_loop = False
# The conf does not need to be dispatch
cfg_id = conf.id
for kind in ('reactionner', 'poller', 'broker',
'receiver'):
r.to_satellites[kind][cfg_id] = None
r.to_satellites_need_dispatch[kind][
cfg_id] = False
r.to_satellites_managed_by[kind][cfg_id] = []
break
logger.info(
'[%s] Trying to send conf %d to scheduler %s' %
(r.get_name(), conf.id, sched.get_name()))
if not sched.need_conf:
logger.info(
'[%s] The scheduler %s do not need conf, sorry'
% (r.get_name(), sched.get_name()))
continue
# We tag conf with the instance_name = scheduler_name
instance_name = sched.scheduler_name
# We give this configuraton a new 'flavor'
conf.push_flavor = random.randint(1, 1000000)
# REF: doc/shinken-conf-dispatching.png (3)
# REF: doc/shinken-scheduler-lost.png (2)
override_conf = sched.get_override_configuration()
satellites_for_sched = r.get_satellites_links_for_scheduler(
)
s_conf = r.serialized_confs[conf.id]
# Prepare the conf before sending it
conf_package = {
'conf': s_conf,
'override_conf': override_conf,
'modules': sched.modules,
'satellites': satellites_for_sched,
'instance_name': sched.scheduler_name,
'push_flavor': conf.push_flavor,
'skip_initial_broks': sched.skip_initial_broks,
}
t1 = time.time()
is_sent = sched.put_conf(conf_package)
logger.debug("Conf is sent in %d" % (time.time() - t1))
if not is_sent:
logger.warning(
'[%s] configuration dispatching error for scheduler %s'
% (r.get_name(), sched.get_name()))
continue
logger.info(
'[%s] Dispatch OK of conf in scheduler %s' %
(r.get_name(), sched.get_name()))
sched.conf = conf
sched.push_flavor = conf.push_flavor
sched.need_conf = False
conf.is_assigned = True
conf.assigned_to = sched
# We update all data for this scheduler
sched.managed_confs = {conf.id: conf.push_flavor}
# Now we generate the conf for satellites:
cfg_id = conf.id
for kind in ('reactionner', 'poller', 'broker',
'receiver'):
r.to_satellites[kind][
cfg_id] = sched.give_satellite_cfg()
r.to_satellites_need_dispatch[kind][cfg_id] = True
r.to_satellites_managed_by[kind][cfg_id] = []
# Ok, the conf is dispatched, no more loop for this
# configuration
break
# We pop conf to dispatch, so it must be no more conf...
conf_to_dispatch = [
cfg for cfg in self.conf.confs.values() if not cfg.is_assigned
]
nb_missed = len(conf_to_dispatch)
if nb_missed > 0:
logger.warning(
"All schedulers configurations are not dispatched, %d are missing"
% nb_missed)
else:
logger.info(
"OK, all schedulers configurations are dispatched :)")
self.dispatch_ok = True
# Sched without conf in a dispatch ok are set to no need_conf
# so they do not raise dispatch where no use
if self.dispatch_ok:
for sched in self.schedulers.items.values():
if sched.conf is None:
# print "Tagging sched", sched.get_name(), "so it do not ask anymore for conf"
sched.need_conf = False
arbiters_cfg = {}
for arb in self.arbiters:
arbiters_cfg[arb.id] = arb.give_satellite_cfg()
# We put the satellites conf with the "new" way so they see only what we want
for r in self.realms:
for cfg in r.confs.values():
cfg_id = cfg.id
# flavor if the push number of this configuration send to a scheduler
flavor = cfg.push_flavor
for kind in ('reactionner', 'poller', 'broker',
'receiver'):
if r.to_satellites_need_dispatch[kind][cfg_id]:
cfg_for_satellite_part = r.to_satellites[kind][
cfg_id]
# make copies of potential_react list for sort
satellites = []
for satellite in r.get_potential_satellites_by_type(
kind):
satellites.append(satellite)
satellites.sort(alive_then_spare_then_deads)
# Only keep alive Satellites and reachable ones
satellites = [
s for s in satellites
if s.alive and s.reachable
]
# If we got a broker, we make the list to pop a new
# item first for each scheduler, so it will smooth the load
# But the spare must stay at the end ;)
if kind == "broker":
nospare = [
s for s in satellites if not s.spare
]
# Should look over the list, not over
if len(nospare) != 0:
idx = cfg_id % len(nospare)
#print "No spare", nospare
spares = [s for s in satellites if s.spare]
#print "Spare", spares
#print "Got 1", nospare[idx:]
#print "Got 2", nospare[:-idx+1]
new_satellites = nospare[idx:]
for _b in nospare[:-idx + 1]:
if _b not in new_satellites:
new_satellites.append(_b)
#new_satellites.extend(nospare[:-idx+1])
#print "New satellites", cfg_id, new_satellites
#for s in new_satellites:
# print "New satellites", cfg_id, s.get_name()
satellites = new_satellites
satellites.extend(spares)
# Dump the order where we will send conf
satellite_string = "[%s] Dispatching %s satellite with order: " % (
r.get_name(), kind)
for satellite in satellites:
satellite_string += '%s (spare:%s), ' % (
satellite.get_name(), str(satellite.spare))
logger.info(satellite_string)
# Now we dispatch cfg to every one ask for it
nb_cfg_sent = 0
for satellite in satellites:
# Send only if we need, and if we can
if nb_cfg_sent < r.get_nb_of_must_have_satellites(
kind) and satellite.alive:
satellite.cfg['schedulers'][
cfg_id] = cfg_for_satellite_part
if satellite.manage_arbiters:
satellite.cfg[
'arbiters'] = arbiters_cfg
# Brokers should have poller/reactionners links too
if kind == "broker":
r.fill_broker_with_poller_reactionner_links(
satellite)
is_sent = False
# Maybe this satellite already got this configuration, so skip it
if satellite.do_i_manage(cfg_id, flavor):
logger.info(
'[%s] Skipping configuration %d send to the %s %s: it already got it'
% (r.get_name(), cfg_id, kind,
satellite.get_name()))
is_sent = True
else: # ok, it really need it :)
logger.info(
'[%s] Trying to send configuration to %s %s'
% (r.get_name(), kind,
satellite.get_name()))
is_sent = satellite.put_conf(
satellite.cfg)
if is_sent:
satellite.active = True
logger.info(
'[%s] Dispatch OK of configuration %s to %s %s'
% (r.get_name(), cfg_id, kind,
satellite.get_name()))
# We change the satellite configuration, update our data
satellite.known_conf_managed_push(
cfg_id, flavor)
nb_cfg_sent += 1
r.to_satellites_managed_by[kind][
cfg_id].append(satellite)
# If we got a broker, the conf_id must be sent to only ONE
# broker, so here it's done, we are happy.
if kind == "broker":
break
#If receiver, we must send the hostnames of this configuration
if kind == 'receiver':
hnames = [
h.get_name() for h in cfg.hosts
]
logger.debug(
"[%s] Sending %s hostnames to the receiver %s"
% (r.get_name(), len(hnames),
satellite.get_name()))
satellite.push_host_names(
cfg_id, hnames)
# else:
# #I've got enough satellite, the next ones are considered spares
if nb_cfg_sent == r.get_nb_of_must_have_satellites(
kind):
logger.info("[%s] OK, no more %s sent need" %
(r.get_name(), kind))
r.to_satellites_need_dispatch[kind][
cfg_id] = False
# And now we dispatch receivers. It's easier, they need ONE conf
# in all their life :)
for r in self.realms:
for rec in r.receivers:
if rec.need_conf:
logger.info(
'[%s] Trying to send configuration to receiver %s'
% (r.get_name(), rec.get_name()))
is_sent = rec.put_conf(rec.cfg)
if is_sent:
rec.active = True
rec.need_conf = False
logger.info(
'[%s] Dispatch OK of configuration to receiver %s'
% (r.get_name(), rec.get_name()))
else:
logger.error(
'[%s] Dispatching failed for receiver %s' %
(r.get_name(), rec.get_name()))
def execute_query(self, query):
"""Just run the query"""
logger.debug("[SqliteDB] Info: I run query '%s'" % query)
self.db_cursor.execute(query)
self.db.commit()
def handle_request_and_fail(self, data):
"""Execute the livestatus request.
This function creates a LiveStatusRequest method, calls the parser,
handles the execution of the request and formatting of the result.
"""
request = LiveStatusRequest(data, self.datamgr, self.query_cache,
self.db, self.pnp_path, self.return_queue,
self.counters)
request.parse_input(data)
if sorted([q.my_type
for q in request.queries]) == ['command', 'query', 'wait']:
# The Multisite way
for query in [
q for q in request.queries if q.my_type == 'command'
]:
result = query.launch_query()
response = query.response
response.format_live_data(result, query.columns, query.aliases)
output, keepalive = response.respond()
output = [q for q in request.queries if q.my_type == 'wait'
] + [q for q in request.queries if q.my_type == 'query']
elif sorted([q.my_type for q in request.queries]) == ['query', 'wait']:
# The Thruk way
output = [q for q in request.queries if q.my_type == 'wait'
] + [q for q in request.queries if q.my_type == 'query']
keepalive = True
elif sorted([q.my_type
for q in request.queries]) == ['command', 'query']:
for query in [
q for q in request.queries if q.my_type == 'command'
]:
result = query.launch_query()
response = query.response
response.format_live_data(result, query.columns, query.aliases)
output, keepalive = response.respond()
for query in [q for q in request.queries if q.my_type == 'query']:
# This was a simple query, respond immediately
result = query.launch_query()
# Now bring the retrieved information to a form which can be sent back to the client
response = query.response
response.format_live_data(result, query.columns, query.aliases)
output, keepalive = response.respond()
elif sorted([q.my_type for q in request.queries]) == ['query']:
for query in [q for q in request.queries if q.my_type == 'query']:
# This was a simple query, respond immediately
result = query.launch_query()
# Now bring the retrieved information to a form which can be sent back to the client
response = query.response
response.format_live_data(result, query.columns, query.aliases)
output, keepalive = response.respond()
elif sorted([q.my_type for q in request.queries]) == ['command']:
for query in [
q for q in request.queries if q.my_type == 'command'
]:
result = query.launch_query()
response = query.response
response.format_live_data(result, query.columns, query.aliases)
output, keepalive = response.respond()
elif [q.my_type for q in request.queries
if q.my_type != 'command'] == []:
# Only external commands. Thruk uses it when it sends multiple
# objects into a downtime.
for query in [
q for q in request.queries if q.my_type == 'command'
]:
result = query.launch_query()
response = query.response
response.format_live_data(result, query.columns, query.aliases)
output, keepalive = response.respond()
else:
# We currently do not handle this kind of composed request
output = ""
logger.error(
"[Livestatus] We currently do not handle this kind of composed request"
)
print sorted([q.my_type for q in request.queries])
logger.debug("[Livestatus] Request duration %.4fs" %
(time.time() - request.tic))
return output, keepalive
def do_publish(to_pack='.'):
logger.debug("WILL CALL PUBLISH.py with %s" % to_pack)
archive = create_archive(to_pack)
publish_archive(archive)
def setup_new_conf(self):
conf = self.new_conf
self.new_conf = None
self.cur_conf = conf
# Got our name from the globals
if 'receiver_name' in conf['global']:
name = conf['global']['receiver_name']
else:
name = 'Unnamed receiver'
self.name = name
self.api_key = conf['global']['api_key']
self.secret = conf['global']['secret']
self.http_proxy = conf['global']['http_proxy']
self.statsd_host = conf['global']['statsd_host']
self.statsd_port = conf['global']['statsd_port']
self.statsd_prefix = conf['global']['statsd_prefix']
self.statsd_enabled = conf['global']['statsd_enabled']
statsmgr.register(self, self.name, 'receiver',
api_key=self.api_key, secret=self.secret, http_proxy=self.http_proxy,
statsd_host=self.statsd_host, statsd_port=self.statsd_port,
statsd_prefix=self.statsd_prefix, statsd_enabled=self.statsd_enabled)
logger.load_obj(self, name)
self.direct_routing = conf['global']['direct_routing']
self.accept_passive_unknown_check_results = \
conf['global']['accept_passive_unknown_check_results']
g_conf = conf['global']
# If we've got something in the schedulers, we do not want it anymore
for sched_id in conf['schedulers']:
already_got = False
# We can already got this conf id, but with another address
if sched_id in self.schedulers:
new_addr = conf['schedulers'][sched_id]['address']
old_addr = self.schedulers[sched_id]['address']
new_port = conf['schedulers'][sched_id]['port']
old_port = self.schedulers[sched_id]['port']
# Should got all the same to be ok :)
if new_addr == old_addr and new_port == old_port:
already_got = True
if already_got:
logger.info("[%s] We already got the conf %d (%s)",
self.name, sched_id, conf['schedulers'][sched_id]['name'])
wait_homerun = self.schedulers[sched_id]['wait_homerun']
actions = self.schedulers[sched_id]['actions']
external_commands = self.schedulers[sched_id]['external_commands']
con = self.schedulers[sched_id]['con']
s = conf['schedulers'][sched_id]
self.schedulers[sched_id] = s
if s['name'] in g_conf['satellitemap']:
s.update(g_conf['satellitemap'][s['name']])
proto = 'http'
if s['use_ssl']:
proto = 'https'
uri = '%s://%s:%s/' % (proto, s['address'], s['port'])
self.schedulers[sched_id]['uri'] = uri
if already_got:
self.schedulers[sched_id]['wait_homerun'] = wait_homerun
self.schedulers[sched_id]['actions'] = actions
self.schedulers[sched_id]['external_commands'] = external_commands
self.schedulers[sched_id]['con'] = con
else:
self.schedulers[sched_id]['wait_homerun'] = {}
self.schedulers[sched_id]['actions'] = {}
self.schedulers[sched_id]['external_commands'] = []
self.schedulers[sched_id]['con'] = None
self.schedulers[sched_id]['running_id'] = 0
self.schedulers[sched_id]['active'] = s['active']
self.schedulers[sched_id]['timeout'] = s['timeout']
self.schedulers[sched_id]['data_timeout'] = s['data_timeout']
# Do not connect if we are a passive satellite
if self.direct_routing and not already_got:
# And then we connect to it :)
self.pynag_con_init(sched_id)
logger.debug("[%s] Sending us configuration %s", self.name, conf)
if not self.have_modules:
self.modules = mods = conf['global']['modules']
self.have_modules = True
logger.info("We received modules %s ", mods)
# Set our giving timezone from arbiter
use_timezone = conf['global']['use_timezone']
if use_timezone != 'NOTSET':
logger.info("Setting our timezone to %s", use_timezone)
os.environ['TZ'] = use_timezone
time.tzset()
def user_login():
logger.debug("[WebUI] user login request, remote user enabled: %s: %s",
app.remote_user_enable, app.remote_user_variable)
for header in app.request.headers:
logger.debug("[WebUI] X request header: %s = %s", header,
app.request.headers[header])
err = app.request.GET.get('error', None)
if err:
logger.warning("[WebUI] login page with error message: %s", err)
cookie_value = app.request.get_cookie(app.session_cookie,
secret=app.auth_secret)
if cookie_value:
logger.info("[WebUI] user login request, existing cookie found: %s",
cookie_value)
# For Alignak backend
if app.alignak_backend_endpoint:
if 'session' in cookie_value:
app.user_session = cookie_value['session']
# For alignak backend
try:
if app.frontend.connect(app.user_session):
bottle.redirect(app.get_url("Dashboard"))
except Exception:
pass
app.response.set_cookie(str(app.session_cookie),
'',
secret=app.auth_secret,
path='/')
bottle.redirect("/user/login")
bottle.redirect(app.get_url("Dashboard"))
elif app.remote_user_enable in ['1', '2']:
logger.debug("[WebUI] user login request, no existing cookie found")
if not err:
user_name = None
if app.remote_user_enable == '1':
logger.debug("[WebUI] search %s in request headers",
app.remote_user_variable)
if app.remote_user_variable in app.request.headers:
user_name = app.request.headers[app.remote_user_variable]
logger.debug(
"[WebUI] remote user found in request headers: %s",
user_name)
if app.remote_user_enable == '2':
logger.debug("[WebUI] search %s in WSGI environment",
app.remote_user_variable)
if app.remote_user_variable in app.request.environ:
user_name = app.request.environ[app.remote_user_variable]
logger.debug(
"[WebUI] remote user found in WSGI environment: %s",
user_name)
if not user_name:
logger.warning(
"[WebUI] remote user is enabled but no authenticated user name was found"
)
bottle.redirect(app.get_url("GetLogin"))
c = app.datamgr.get_contact(name=user_name)
if c:
cookie_value = {
'login': user_name,
'session': app.user_session,
'info': app.user_info
}
app.response.set_cookie(str(app.session_cookie),
cookie_value,
secret=app.auth_secret,
path='/')
bottle.redirect(app.get_url("Dashboard"))
return {
'msg_text': err,
'login_text': app.login_text,
'company_logo': app.company_logo,
'path': app.request.query.get('path', '')
}
def install_package(pname, raw):
logger.debug("Installing the package %s (size:%d)" % (pname, len(raw)))
if len(raw) == 0:
logger.error('The package %s cannot be found' % pname)
return
tmpdir = os.path.join(tempfile.gettempdir(), pname)
logger.debug("Unpacking the package into %s" % tmpdir)
if os.path.exists(tmpdir):
logger.debug("Removing previous tmp dir %s" % tmpdir)
shutil.rmtree(tmpdir)
logger.debug("Creating temporary dir %s" % tmpdir)
os.mkdir(tmpdir)
package_content = []
# open a file with the content
f = StringIO(raw)
tar_file = tarfile.open(fileobj=f, mode="r")
logger.debug("Tar file contents:")
for i in tar_file.getmembers():
path = i.name
if path == '.':
continue
if path.startswith('/') or '..' in path:
logger.error("SECURITY: the path %s seems dangerous!" % path)
return
# Adding all files into the package_content list
package_content.append({
'name': i.name,
'mode': i.mode,
'type': i.type,
'size': i.size
})
logger.debug("\t%s" % path)
# Extract all in the tmpdir
tar_file.extractall(tmpdir)
tar_file.close()
# Now we look at the package.json that will give us our name and co
package_json_p = os.path.join(tmpdir, 'package.json')
if not os.path.exists(package_json_p):
logger.error("Error : bad archive : Missing file %s" % package_json_p)
return None
package_json = read_package_json(open(package_json_p))
logger.debug("Package.json content %s " % package_json)
modules_dir = CONFIG['paths']['modules']
share_dir = CONFIG['paths']['share']
packs_dir = CONFIG['paths']['packs']
etc_dir = CONFIG['paths']['etc']
doc_dir = CONFIG['paths']['doc']
inventory_dir = CONFIG['paths']['inventory']
libexec_dir = CONFIG['paths'].get(
'libexec', os.path.join(CONFIG['paths']['lib'], 'libexec'))
test_dir = CONFIG['paths'].get('test', '/__DONOTEXISTS__')
for d in (modules_dir, share_dir, packs_dir, doc_dir, inventory_dir):
if not os.path.exists(d):
logger.error("The installation directory %s is missing!" % d)
return
# Now install the package from $TMP$/share/* to $SHARE$/*
p_share = os.path.join(tmpdir, 'share')
logger.debug("TMPDIR:%s aahre_dir:%s pname:%s" %
(tmpdir, share_dir, pname))
if os.path.exists(p_share):
logger.info("Installing the share package data")
# shutil will do the create dir
_copytree(p_share, share_dir)
logger.info("Copy done in the share directory %s" % share_dir)
logger.debug("TMPDIR:%s modules_dir:%s pname:%s" %
(tmpdir, modules_dir, pname))
# Now install the package from $TMP$/module/* to $MODULES$/pname/*
p_module = os.path.join(tmpdir, 'module')
if os.path.exists(p_module):
logger.info("Installing the module package data")
mod_dest = os.path.join(modules_dir, pname)
if os.path.exists(mod_dest):
logger.info("Removing previous module install at %s" % mod_dest)
shutil.rmtree(mod_dest)
# shutil will do the create dir
shutil.copytree(p_module, mod_dest)
logger.info("Copy done in the module directory %s" % mod_dest)
p_doc = os.path.join(tmpdir, 'doc')
logger.debug("TMPDIR:%s doc_dir:%s pname:%s" % (tmpdir, doc_dir, pname))
# Now install the package from $TMP$/doc/* to $MODULES$/doc/source/89_packages/pname/*
if os.path.exists(p_doc):
logger.info("Installing the doc package data")
doc_dest = os.path.join(doc_dir, 'source', '89_packages', pname)
if os.path.exists(doc_dest):
logger.info("Removing previous doc install at %s" % doc_dest)
shutil.rmtree(doc_dest)
# shutil will do the create dir
shutil.copytree(p_doc, doc_dest)
logger.info("Copy done in the doc directory %s" % doc_dest)
# Now install the pack from $TMP$/pack/* to $PACKS$/pname/*
p_pack = os.path.join(tmpdir, 'pack')
if os.path.exists(p_pack):
logger.info("Installing the pack package data")
pack_dest = os.path.join(packs_dir, pname)
if os.path.exists(pack_dest):
logger.info("Removing previous pack install at %s" % pack_dest)
shutil.rmtree(pack_dest)
# shutil will do the create dir
shutil.copytree(p_pack, pack_dest)
logger.info("Copy done in the pack directory %s" % pack_dest)
# Now install the etc from $TMP$/etc/* to $ETC$/etc/*
p_etc = os.path.join(tmpdir, 'etc')
if os.path.exists(p_etc):
logger.info("Merging the etc package data into your etc directory")
# We don't use shutils because it NEED etc_dir to be non existant...
# Come one guys..... cp is not as terrible as this...
_copytree(p_etc, etc_dir)
logger.info("Copy done in the etc directory %s" % etc_dir)
# Now install the tests from $TMP$/tests/* to $TESTS$/tests/*
# if the last one is specified on the configuration file (optionnal)
p_tests = os.path.join(tmpdir, 'test')
if os.path.exists(p_tests) and os.path.exists(test_dir):
logger.info("Merging the test package data into your test directory")
# We don't use shutils because it NEED etc_dir to be non existant...
# Come one guys..... cp is not as terrible as this...
logger.debug("COPYING %s into %s" % (p_tests, test_dir))
_copytree(p_tests, test_dir)
logger.info("Copy done in the test directory %s" % test_dir)
# Now install the libexec things from $TMP$/libexec/* to $LIBEXEC$/*
# but also chmod a+x the plugins copied
p_libexec = os.path.join(tmpdir, 'libexec')
if os.path.exists(p_libexec) and os.path.exists(libexec_dir):
logger.info(
"Merging the libexec package data into your libexec directory")
logger.debug("COPYING %s into %s" % (p_libexec, libexec_dir))
# Before be sure all files in there are +x
_chmodplusx(p_libexec)
_copytree(p_libexec, libexec_dir)
logger.info("Copy done in the libexec directory %s" % libexec_dir)
# then samve the package.json into the inventory dir
p_inv = os.path.join(inventory_dir, pname)
if not os.path.exists(p_inv):
os.mkdir(p_inv)
shutil.copy2(package_json_p, os.path.join(p_inv, 'package.json'))
# and the package content
cont = open(os.path.join(p_inv, 'content.json'), 'w')
cont.write(json.dumps(package_content))
cont.close()
# We now clean (rm) the tmpdir we don't need any more
try:
shutil.rmtree(tmpdir, ignore_errors=True)
# cannot remove? not a crime
except OSError:
pass
# THE END, output all is OK :D
cprint('OK ', 'green', end='')
cprint('%s' % pname)
def push_external_commands_to_schedulers(self):
# If we are not in a direct routing mode, just bailout after
# faking resolving the commands
if not self.direct_routing:
self.external_commands.extend(self.unprocessed_external_commands)
self.unprocessed_external_commands = []
return
# Now get all external commands and put them into the
# good schedulers
for ext_cmd in self.unprocessed_external_commands:
self.external_command.resolve_command(ext_cmd)
self.external_commands.append(ext_cmd)
# And clean the previous one
self.unprocessed_external_commands = []
# Now for all alive schedulers, send the commands
for sched_id in self.schedulers:
sched = self.schedulers[sched_id]
extcmds = sched['external_commands']
cmds = [extcmd.cmd_line for extcmd in extcmds]
con = sched.get('con', None)
sent = False
if not con:
logger.warning("The scheduler is not connected" % sched)
self.pynag_con_init(sched_id)
con = sched.get('con', None)
# If there are commands and the scheduler is alive
if len(cmds) > 0 and con:
logger.debug("Sending %d commands to scheduler %s" %
(len(cmds), sched))
try:
con.run_external_commands(cmds)
sent = True
# Not connected or sched is gone
except (Pyro_exp_pack, KeyError), exp:
logger.debug('manage_returns exception:: %s,%s ' %
(type(exp), str(exp)))
try:
logger.debug(
''.join(PYRO_VERSION < "4.0"
and Pyro.util.getPyroTraceback(exp)
or Pyro.util.getPyroTraceback()))
except:
pass
self.pynag_con_init(sched_id)
return
except AttributeError, exp: # the scheduler must not be initialized
logger.debug('manage_returns exception:: %s,%s ' %
(type(exp), str(exp)))
except Exception, exp:
logger.error(
"A satellite raised an unknown exception: %s (%s)" %
(exp, type(exp)))
try:
logger.debug(
''.join(PYRO_VERSION < "4.0"
and Pyro.util.getPyroTraceback(exp)
or Pyro.util.getPyroTraceback()))
except:
pass
raise
c.setopt(pycurl.WRITEFUNCTION, response.write)
#c.setopt(c.VERBOSE, 1)
try:
c.perform()
except pycurl.error, exp:
logger.error("There was a critical error : %s" % exp)
return ''
r = c.getinfo(pycurl.HTTP_CODE)
c.close()
if r != 200:
logger.error("There was a critical error : %s" % response.getvalue())
sys.exit(2)
else:
ret = response.getvalue()
logger.debug("CURL result len : %d " % len(ret))
return ret
def grab_local(d):
# First try to look if the directory we are trying to pack is valid
to_pack = os.path.abspath(d)
if not os.path.exists(to_pack):
err = "Error : the directory to install is missing %s" % to_pack
logger.error(err)
raise Exception(err)
package_json_p = os.path.join(to_pack, 'package.json')
if not os.path.exists(package_json_p):
logger.error("Error : Missing file %s" % package_json_p)
sys.exit(2)
def wait_new_conf(self):
logger.debug("Arbiter wants me to wait for a new configuration")
self.app.sched.die()
super(IForArbiter, self).wait_new_conf()
def __init__(self,
sqlite_cursor=None,
sqlite_row=None,
line=None,
srcdict=None):
if srcdict != None:
for col in Logline.columns:
logger.info("[Livestatus Log Lines] Set %s, %s" %
(col, srcdict[col]))
setattr(self, col, srcdict[col])
elif sqlite_cursor != None and sqlite_row != None:
for idx, col in enumerate(sqlite_cursor):
if col[0] == 'class':
setattr(self, 'logclass', sqlite_row[idx])
else:
setattr(self, col[0], sqlite_row[idx])
elif line != None:
if isinstance(line, unicode):
line = line.encode('UTF-8').rstrip()
# [1278280765] SERVICE ALERT: test_host_0
if line[0] != '[' and line[11] != ']':
logger.warning("[Livestatus Log Lines] Invalid line: %s" %
line)
raise LoglineWrongFormat
else:
service_states = {
'OK': 0,
'WARNING': 1,
'CRITICAL': 2,
'UNKNOWN': 3,
'RECOVERY': 0
}
host_states = {
'UP': 0,
'DOWN': 1,
'UNREACHABLE': 2,
'UNKNOWN': 3,
'RECOVERY': 0
}
# type is 0:info, 1:state, 2:program, 3:notification, 4:passive, 5:command
logobject = LOGOBJECT_INFO
logclass = LOGCLASS_INVALID
attempt, state = [0] * 2
command_name, comment, contact_name, host_name, message, plugin_output, service_description, state_type = [
''
] * 8
time = line[1:11]
first_type_pos = line.find(' ') + 1
last_type_pos = line.find(':')
first_detail_pos = last_type_pos + 2
type = line[first_type_pos:last_type_pos]
options = line[first_detail_pos:]
message = line
if type == 'CURRENT SERVICE STATE':
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_STATE
host_name, service_description, state, state_type, attempt, plugin_output = options.split(
';', 5)
elif type == 'INITIAL SERVICE STATE':
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_STATE
host_name, service_description, state, state_type, attempt, plugin_output = options.split(
';', 5)
elif type == 'SERVICE ALERT':
# SERVICE ALERT: srv-40;Service-9;CRITICAL;HARD;1;[Errno 2] No such file or directory
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_ALERT
host_name, service_description, state, state_type, attempt, plugin_output = options.split(
';', 5)
state = service_states[state]
elif type == 'SERVICE DOWNTIME ALERT':
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_ALERT
host_name, service_description, state_type, comment = options.split(
';', 3)
elif type == 'SERVICE FLAPPING ALERT':
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_ALERT
host_name, service_description, state_type, comment = options.split(
';', 3)
elif type == 'CURRENT HOST STATE':
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_STATE
host_name, state, state_type, attempt, plugin_output = options.split(
';', 4)
elif type == 'INITIAL HOST STATE':
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_STATE
host_name, state, state_type, attempt, plugin_output = options.split(
';', 4)
elif type == 'HOST ALERT':
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_ALERT
host_name, state, state_type, attempt, plugin_output = options.split(
';', 4)
state = host_states[state]
elif type == 'HOST DOWNTIME ALERT':
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_ALERT
host_name, state_type, comment = options.split(';', 2)
elif type == 'HOST FLAPPING ALERT':
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_ALERT
host_name, state_type, comment = options.split(';', 2)
elif type == 'SERVICE NOTIFICATION':
# tust_cuntuct;test_host_0;test_ok_0;CRITICAL;notify-service;i am CRITICAL <-- normal
# SERVICE NOTIFICATION: test_contact;test_host_0;test_ok_0;DOWNTIMESTART (OK);notify-service;OK
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_NOTIFICATION
contact_name, host_name, service_description, state_type, command_name, check_plugin_output = options.split(
';', 5)
if '(' in state_type: # downtime/flapping/etc-notifications take the type UNKNOWN
state_type = 'UNKNOWN'
state = service_states[state_type]
elif type == 'HOST NOTIFICATION':
# tust_cuntuct;test_host_0;DOWN;notify-host;i am DOWN
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_NOTIFICATION
contact_name, host_name, state_type, command_name, check_plugin_output = options.split(
';', 4)
if '(' in state_type:
state_type = 'UNKNOWN'
state = host_states[state_type]
elif type == 'PASSIVE SERVICE CHECK':
logobject = LOGOBJECT_SERVICE
logclass = LOGCLASS_PASSIVECHECK
host_name, service_description, state, check_plugin_output = options.split(
';', 3)
elif type == 'PASSIVE HOST CHECK':
logobject = LOGOBJECT_HOST
logclass = LOGCLASS_PASSIVECHECK
host_name, state, check_plugin_output = options.split(
';', 2)
elif type == 'SERVICE EVENT HANDLER':
# SERVICE EVENT HANDLER: test_host_0;test_ok_0;CRITICAL;SOFT;1;eventhandler
logobject = LOGOBJECT_SERVICE
host_name, service_description, state, state_type, attempt, command_name = options.split(
';', 5)
state = service_states[state]
elif type == 'HOST EVENT HANDLER':
logobject = LOGOBJECT_HOST
host_name, state, state_type, attempt, command_name = options.split(
';', 4)
state = host_states[state]
elif type == 'EXTERNAL COMMAND':
logobject = LOGOBJECT_INFO
logclass = LOGCLASS_COMMAND
elif type.startswith('starting...') or \
type.startswith('shutting down...') or \
type.startswith('Bailing out') or \
type.startswith('active mode...') or \
type.startswith('standby mode...') or \
type.startswith('Warning'):
logobject = LOGOBJECT_INFO
logclass = LOGCLASS_PROGRAM
else:
logger.debug("[Livestatus Log Lines] Does not match")
pass
Logline.id += 1
self.lineno = Logline.id
setattr(self, 'logobject', int(logobject))
setattr(self, 'attempt', int(attempt))
setattr(self, 'logclass', int(logclass))
setattr(self, 'command_name', command_name)
setattr(self, 'comment', comment)
setattr(self, 'contact_name', contact_name)
setattr(self, 'host_name', host_name)
setattr(self, 'message', message)
setattr(
self, 'options', ''
) # Fix a mismatch of number of fields with old databases and new ones
setattr(self, 'plugin_output', plugin_output)
setattr(self, 'service_description', service_description)
setattr(self, 'state', state)
setattr(self, 'state_type', state_type)
setattr(self, 'time', int(time))
setattr(self, 'type', type)
def setup_new_conf(self):
conf = self.new_conf
self.new_conf = None
self.cur_conf = conf
# Got our name from the globals
g_conf = conf['global']
if 'broker_name' in g_conf:
name = g_conf['broker_name']
else:
name = 'Unnamed broker'
self.name = name
logger.load_obj(self, name)
logger.debug("[%s] Sending us configuration %s" % (self.name, conf))
# If we've got something in the schedulers, we do not
# want it anymore
# self.schedulers.clear()
for sched_id in conf['schedulers']:
# Must look if we already have it to do not overdie our broks
already_got = False
# We can already got this conf id, but with another address
if sched_id in self.schedulers:
new_addr = conf['schedulers'][sched_id]['address']
old_addr = self.schedulers[sched_id]['address']
new_port = conf['schedulers'][sched_id]['port']
old_port = self.schedulers[sched_id]['port']
# Should got all the same to be ok :)
if new_addr == old_addr and new_port == old_port:
already_got = True
if already_got:
broks = self.schedulers[sched_id]['broks']
running_id = self.schedulers[sched_id]['running_id']
else:
broks = {}
running_id = 0
s = conf['schedulers'][sched_id]
self.schedulers[sched_id] = s
# replacing scheduler address and port by those defined in satellitemap
if s['name'] in g_conf['satellitemap']:
s = dict(s) # make a copy
s.update(g_conf['satellitemap'][s['name']])
uri = pyro.create_uri(s['address'], s['port'], 'Broks',
self.use_ssl)
self.schedulers[sched_id]['uri'] = uri
self.schedulers[sched_id]['broks'] = broks
self.schedulers[sched_id]['instance_id'] = s['instance_id']
self.schedulers[sched_id]['running_id'] = running_id
self.schedulers[sched_id]['active'] = s['active']
self.schedulers[sched_id]['last_connection'] = 0
logger.info("We have our schedulers: %s " % self.schedulers)
# Now get arbiter
for arb_id in conf['arbiters']:
# Must look if we already have it
already_got = arb_id in self.arbiters
if already_got:
broks = self.arbiters[arb_id]['broks']
else:
broks = {}
a = conf['arbiters'][arb_id]
self.arbiters[arb_id] = a
# replacing arbiter address and port by those defined in satellitemap
if a['name'] in g_conf['satellitemap']:
a = dict(a) # make a copy
a.update(g_conf['satellitemap'][a['name']])
uri = pyro.create_uri(a['address'], a['port'], 'Broks',
self.use_ssl)
self.arbiters[arb_id]['uri'] = uri
self.arbiters[arb_id]['broks'] = broks
self.arbiters[arb_id]['instance_id'] = 0 # No use so all to 0
self.arbiters[arb_id]['running_id'] = 0
self.arbiters[arb_id]['last_connection'] = 0
# We do not connect to the arbiter. Connection hangs
logger.info("We have our arbiters: %s " % self.arbiters)
# Now for pollers
for pol_id in conf['pollers']:
# Must look if we already have it
already_got = pol_id in self.pollers
if already_got:
broks = self.pollers[pol_id]['broks']
running_id = self.schedulers[sched_id]['running_id']
else:
broks = {}
running_id = 0
p = conf['pollers'][pol_id]
self.pollers[pol_id] = p
# replacing poller address and port by those defined in satellitemap
if p['name'] in g_conf['satellitemap']:
p = dict(p) # make a copy
p.update(g_conf['satellitemap'][p['name']])
uri = pyro.create_uri(p['address'], p['port'], 'Broks',
self.use_ssl)
self.pollers[pol_id]['uri'] = uri
self.pollers[pol_id]['broks'] = broks
self.pollers[pol_id]['instance_id'] = 0 # No use so all to 0
self.pollers[pol_id]['running_id'] = running_id
self.pollers[pol_id]['last_connection'] = 0
# #And we connect to it
# self.app.pynag_con_init(pol_id, 'poller')
logger.info("We have our pollers: %s" % self.pollers)
# Now reactionners
for rea_id in conf['reactionners']:
# Must look if we already have it
already_got = rea_id in self.reactionners
if already_got:
broks = self.reactionners[rea_id]['broks']
running_id = self.schedulers[sched_id]['running_id']
else:
broks = {}
running_id = 0
r = conf['reactionners'][rea_id]
self.reactionners[rea_id] = r
# replacing reactionner address and port by those defined in satellitemap
if r['name'] in g_conf['satellitemap']:
r = dict(r) # make a copy
r.update(g_conf['satellitemap'][r['name']])
uri = pyro.create_uri(r['address'], r['port'], 'Broks',
self.use_ssl)
self.reactionners[rea_id]['uri'] = uri
self.reactionners[rea_id]['broks'] = broks
self.reactionners[rea_id]['instance_id'] = 0 # No use so all to 0
self.reactionners[rea_id]['running_id'] = running_id
self.reactionners[rea_id]['last_connection'] = 0
# #And we connect to it
# self.app.pynag_con_init(rea_id, 'reactionner')
logger.info("We have our reactionners: %s" % self.reactionners)
if not self.have_modules:
self.modules = mods = conf['global']['modules']
self.have_modules = True
logger.info("We received modules %s " % mods)
# Ok now start, or restart them!
# Set modules, init them and start external ones
self.modules_manager.set_modules(self.modules)
self.do_load_modules()
self.modules_manager.start_external_instances()
# Set our giving timezone from arbiter
use_timezone = conf['global']['use_timezone']
if use_timezone != 'NOTSET':
logger.info("Setting our timezone to %s" % use_timezone)
os.environ['TZ'] = use_timezone
time.tzset()
# Connection init with Schedulers
for sched_id in self.schedulers:
self.pynag_con_init(sched_id, type='scheduler')
for pol_id in self.pollers:
self.pynag_con_init(pol_id, type='poller')
for rea_id in self.reactionners:
self.pynag_con_init(rea_id, type='reactionner')
def setup_new_conf(self):
pk = self.new_conf
conf_raw = pk['conf']
override_conf = pk['override_conf']
modules = pk['modules']
satellites = pk['satellites']
instance_name = pk['instance_name']
push_flavor = pk['push_flavor']
skip_initial_broks = pk['skip_initial_broks']
t0 = time.time()
conf = cPickle.loads(conf_raw)
logger.debug("Conf received at %d. Unserialized in %d secs" %
(t0, time.time() - t0))
self.new_conf = None
# Tag the conf with our data
self.conf = conf
self.conf.push_flavor = push_flavor
self.conf.instance_name = instance_name
self.conf.skip_initial_broks = skip_initial_broks
self.cur_conf = conf
self.override_conf = override_conf
self.modules = modules
self.satellites = satellites
#self.pollers = self.app.pollers
if self.conf.human_timestamp_log:
logger.set_human_format()
# Now We create our pollers
for pol_id in satellites['pollers']:
# Must look if we already have it
already_got = pol_id in self.pollers
p = satellites['pollers'][pol_id]
self.pollers[pol_id] = p
if p['name'] in override_conf['satellitemap']:
p = dict(p) # make a copy
p.update(override_conf['satellitemap'][p['name']])
uri = pyro.create_uri(p['address'], p['port'], 'Schedulers',
self.use_ssl)
self.pollers[pol_id]['uri'] = uri
self.pollers[pol_id]['last_connection'] = 0
# First mix conf and override_conf to have our definitive conf
for prop in self.override_conf:
#print "Overriding the property %s with value %s" % (prop, self.override_conf[prop])
val = self.override_conf[prop]
setattr(self.conf, prop, val)
if self.conf.use_timezone != '':
logger.debug("Setting our timezone to %s" %
str(self.conf.use_timezone))
os.environ['TZ'] = self.conf.use_timezone
time.tzset()
if len(self.modules) != 0:
logger.debug("I've got %s modules" % str(self.modules))
# TODO: if scheduler had previous modules instanciated it must clean them!
self.modules_manager.set_modules(self.modules)
self.do_load_modules()
# give it an interface
# But first remove previous interface if exists
if self.ichecks is not None:
logger.debug(
"Deconnecting previous Check Interface from pyro_daemon")
self.pyro_daemon.unregister(self.ichecks)
# Now create and connect it
self.ichecks = IChecks(self.sched)
self.uri = self.pyro_daemon.register(self.ichecks, "Checks")
logger.debug("The Checks Interface uri is: %s" % self.uri)
# Same for Broks
if self.ibroks is not None:
logger.debug(
"Deconnecting previous Broks Interface from pyro_daemon")
self.pyro_daemon.unregister(self.ibroks)
# Create and connect it
self.ibroks = IBroks(self.sched)
self.uri2 = self.pyro_daemon.register(self.ibroks, "Broks")
logger.debug("The Broks Interface uri is: %s" % self.uri2)
logger.info("Loading configuration.")
self.conf.explode_global_conf()
# we give sched it's conf
self.sched.reset()
self.sched.load_conf(self.conf)
self.sched.load_satellites(self.pollers, self.reactionners)
# We must update our Config dict macro with good value
# from the config parameters
self.sched.conf.fill_resource_macros_names_macros()
#print "DBG: got macros", self.sched.conf.macros
# Creating the Macroresolver Class & unique instance
m = MacroResolver()
m.init(self.conf)
#self.conf.dump()
#self.conf.quick_debug()
# Now create the external commander
# it's a applyer: it role is not to dispatch commands,
# but to apply them
e = ExternalCommandManager(self.conf, 'applyer')
# Scheduler need to know about external command to
# activate it if necessary
self.sched.load_external_command(e)
# External command need the sched because he can raise checks
e.load_scheduler(self.sched)
# We clear our schedulers managed (it's us :) )
# and set ourself in it
self.schedulers = {self.conf.instance_id: self.sched}
class Ip_Tag_Arbiter(BaseModule):
def __init__(self,
mod_conf,
ip_range,
prop,
value,
method,
ignore_hosts=None):
BaseModule.__init__(self, mod_conf)
self.ip_range = IP(ip_range)
self.property = prop
self.value = value
self.method = method
if ignore_hosts:
self.ignore_hosts = ignore_hosts.split(', ')
logger.debug("[IP Tag] Ignoring hosts : %s" % self.ignore_hosts)
else:
self.ignore_hosts = []
self.pool_size = int(getattr(mod_conf, 'pool_size', '1'))
# Called by Arbiter to say 'let's prepare yourself guy'
def init(self):
logger.info("[IP Tag] Initialization of the ip range tagger module")
def hook_early_configuration(self, arb):
logger.info("[IpTag] in hook late config")
# Get a pool for gevent jobs
if Pool:
pool = Pool(100)
else:
pool = None
for h in arb.conf.hosts:
if not hasattr(h, 'address') and not hasattr(h, 'host_name'):
continue
if h.get_name() in self.ignore_hosts:
logger.debug("[IP Tag] Ignoring host %s" % h.get_name())
continue
# The address to resolve
addr = None
# By default take the address, if not, take host_name
if not hasattr(h, 'address'):
addr = h.host_name
else:
addr = h.address
logger.debug("[IP Tag] Looking for %s" % h.get_name())
logger.debug("[IP Tag] Address is %s" % str(addr))
h_ip = None
try:
IP(addr)
# If we reach here, it's it was a real IP :)
h_ip = addr
except:
pass
if pool:
pool.spawn(self.job, h, h_ip, addr)
else:
self.job(h, h_ip, addr)
# Now wait for all jobs to finish if need
if pool:
pool.join()
# Main job, will manage eachhost in asyncronous mode thanks to gevent
def job(self, h, h_ip, addr):
# Ok, try again with name resolution
if not h_ip:
try:
h_ip = socket.gethostbyname(addr)
except Exception, exp:
pass
# Ok, maybe we succeed :)
logger.debug("[IP Tag] Host ip is: %s" % str(h_ip))
# If we got an ip that match and the object do not already got
# the property, tag it!
if h_ip and h_ip in self.ip_range:
logger.debug("[IP Tag] Is in the range")
# 4 cases: append , replace and set
# append will join with the value if exist (on the END)
# prepend will join with the value if exist (on the BEGINING)
# replace will replace it if NOT existing
# set put the value even if the property exists
if self.method == 'append':
orig_v = getattr(h, self.property, '')
new_v = ','.join([orig_v, self.value])
setattr(h, self.property, new_v)
# Same but we put before
if self.method == 'prepend':
orig_v = getattr(h, self.property, '')
new_v = ','.join([self.value, orig_v])
setattr(h, self.property, new_v)
if self.method == 'replace':
if not hasattr(h, self.property):
# Ok, set the value!
setattr(h, self.property, self.value)
if self.method == 'set':
setattr(h, self.property, self.value)
class Broker(BaseSatellite):
properties = BaseSatellite.properties.copy()
properties.update({
'pidfile': PathProp(default='brokerd.pid'),
'port': IntegerProp(default='7772'),
'local_log': PathProp(default='brokerd.log'),
})
def __init__(self, config_file, is_daemon, do_replace, debug, debug_file):
super(Broker, self).__init__('broker', config_file, is_daemon,
do_replace, debug, debug_file)
# Our arbiters
self.arbiters = {}
# Our pollers and reactionners
self.pollers = {}
self.reactionners = {}
# Modules are load one time
self.have_modules = False
# Can have a queue of external_commands given by modules
# will be processed by arbiter
self.external_commands = []
# All broks to manage
self.broks = [] # broks to manage
# broks raised this turn and that needs to be put in self.broks
self.broks_internal_raised = []
self.timeout = 1.0
# Schedulers have some queues. We can simplify the call by adding
# elements into the proper queue just by looking at their type
# Brok -> self.broks
# TODO: better tag ID?
# External commands -> self.external_commands
def add(self, elt):
cls_type = elt.__class__.my_type
if cls_type == 'brok':
# For brok, we TAG brok with our instance_id
elt.instance_id = 0
self.broks_internal_raised.append(elt)
return
elif cls_type == 'externalcommand':
logger.debug("Enqueuing an external command '%s'" %
str(ExternalCommand.__dict__))
self.external_commands.append(elt)
# Maybe we got a Message from the modules, it's way to ask something
# like from now a full data from a scheduler for example.
elif cls_type == 'message':
# We got a message, great!
logger.debug(str(elt.__dict__))
if elt.get_type() == 'NeedData':
data = elt.get_data()
# Full instance id means: I got no data for this scheduler
# so give me all dumbass!
if 'full_instance_id' in data:
c_id = data['full_instance_id']
source = elt.source
logger.info(
'The module %s is asking me to get all initial data from the scheduler %d'
% (source, c_id))
# so we just reset the connection and the running_id, it will just get all new things
try:
self.schedulers[c_id]['con'] = None
self.schedulers[c_id]['running_id'] = 0
except KeyError: # maybe this instance was not known, forget it
logger.warning(
"the module %s ask me a full_instance_id for an unknown ID (%d)!"
% (source, c_id))
# Maybe a module tells me that it's dead, I must log it's last words...
if elt.get_type() == 'ICrash':
data = elt.get_data()
logger.error(
'the module %s just crash! Please look at the traceback:' %
data['name'])
logger.error(data['trace'])
# The module death will be looked for elsewhere and restarted.
# Get the good tabs for links by the kind. If unknown, return None
def get_links_from_type(self, type):
t = {'scheduler': self.schedulers, 'arbiter': self.arbiters, \
'poller': self.pollers, 'reactionner': self.reactionners}
if type in t:
return t[type]
return None
# Call by arbiter to get our external commands
def get_external_commands(self):
res = self.external_commands
self.external_commands = []
return res
# Check if we do not connect to often to this
def is_connection_try_too_close(self, elt):
now = time.time()
last_connection = elt['last_connection']
if now - last_connection < 5:
return True
return False
# initialize or re-initialize connection with scheduler or
# arbiter if type == arbiter
def pynag_con_init(self, id, type='scheduler'):
# Get the good links tab for looping..
links = self.get_links_from_type(type)
if links is None:
logger.debug('Type unknown for connection! %s' % type)
return
if type == 'scheduler':
# If sched is not active, I do not try to init
# it is just useless
is_active = links[id]['active']
if not is_active:
return
# If we try to connect too much, we slow down our tests
if self.is_connection_try_too_close(links[id]):
return
# Ok, we can now update it
links[id]['last_connection'] = time.time()
# DBG: print "Init connection with", links[id]['uri']
running_id = links[id]['running_id']
# DBG: print "Running id before connection", running_id
uri = links[id]['uri']
try:
socket.setdefaulttimeout(3)
links[id]['con'] = Pyro.core.getProxyForURI(uri)
socket.setdefaulttimeout(None)
except Pyro_exp_pack, exp:
# But the multiprocessing module is not compatible with it!
# so we must disable it immediately after
socket.setdefaulttimeout(None)
logger.info("Connection problem to the %s %s: %s" %
(type, links[id]['name'], str(exp)))
links[id]['con'] = None
return
try:
# initial ping must be quick
pyro.set_timeout(links[id]['con'], 5)
links[id]['con'].ping()
new_run_id = links[id]['con'].get_running_id()
# data transfer can be longer
pyro.set_timeout(links[id]['con'], 120)
# The schedulers have been restarted: it has a new run_id.
# So we clear all verifs, they are obsolete now.
if new_run_id != running_id:
logger.debug("[%s] New running id for the %s %s: %s (was %s)" %
(self.name, type, links[id]['name'], new_run_id,
running_id))
links[id]['broks'].clear()
# we must ask for a new full broks if
# it's a scheduler
if type == 'scheduler':
logger.debug(
"[%s] I ask for a broks generation to the scheduler %s"
% (self.name, links[id]['name']))
links[id]['con'].fill_initial_broks(self.name)
# Ok all is done, we can save this new running id
links[id]['running_id'] = new_run_id
except Pyro_exp_pack, exp:
logger.info("Connection problem to the %s %s: %s" %
(type, links[id]['name'], str(exp)))
links[id]['con'] = None
return
def get_graph_uris(self, elt, graphstart, graphend, source='detail'):
# Ugly to hard-code such values. But where else should I put them ?
fontsize = {'detail': '8', 'dashboard': '18'}
if not elt:
return []
t = elt.__class__.my_type
r = []
# Hanling Graphite variables
data_source = ""
graphite_pre = ""
graphite_post = ""
if self.graphite_data_source:
data_source = ".%s" % self.graphite_data_source
if t == 'host':
if "_GRAPHITE_PRE" in elt.customs:
graphite_pre = "%s." % elt.customs["_GRAPHITE_PRE"]
elif t == 'service':
if "_GRAPHITE_PRE" in elt.host.customs:
graphite_pre = "%s." % elt.host.customs["_GRAPHITE_PRE"]
if "_GRAPHITE_POST" in elt.customs:
graphite_post = ".%s" % elt.customs["_GRAPHITE_POST"]
# Format the start & end time (and not only the date)
d = datetime.fromtimestamp(graphstart)
d = d.strftime('%H:%M_%Y%m%d')
e = datetime.fromtimestamp(graphend)
e = e.strftime('%H:%M_%Y%m%d')
filename = elt.check_command.get_name().split('!')[0] + '.graph'
# Do we have a template for the given source?
thefile = os.path.join(self.templates_path, source, filename)
# If not try to use the one for the parent folder
if not os.path.isfile(thefile):
# In case of CHECK_NRPE, the check_name is in second place
if len(elt.check_command.get_name().split('!')) > 1:
filename = elt.check_command.get_name().split(
'!')[0] + '_' + elt.check_command.get_name().split(
'!')[1] + '.graph'
thefile = os.path.join(self.templates_path, source, filename)
if not os.path.isfile(thefile):
thefile = os.path.join(self.templates_path, filename)
logger.debug("[ui-graphite] template=%s", thefile)
if os.path.isfile(thefile):
template_html = ''
with open(thefile, 'r') as template_file:
template_html += template_file.read()
# Read the template file, as template string python object
html = Template(template_html)
# Build the dict to instantiate the template string
values = {}
if t == 'host':
values['host'] = graphite_pre + self.illegal_char.sub(
"_", elt.host_name) + data_source
values['service'] = '__HOST__'
if t == 'service':
values['host'] = graphite_pre + self.illegal_char.sub(
"_", elt.host.host_name) + data_source
values['service'] = self.illegal_char.sub(
"_", elt.service_description) + graphite_post
values['uri'] = self.uri
# Split, we may have several images.
for img in html.substitute(values).split('\n'):
if not img == "":
v = {}
v['link'] = self.uri
v['img_src'] = img.replace(
'"', "'") + "&from=" + d + "&until=" + e
v['img_src'] = self._replaceFontSize(
v['img_src'], fontsize[source])
r.append(v)
# No need to continue, we have the images already.
return r
# If no template is present, then the usual way
if t == 'host':
couples = self.get_metric_and_value(elt.perf_data)
# If no values, we can exit now
if len(couples) == 0:
return []
# Remove all non alpha numeric character
host_name = self.illegal_char.sub('_', elt.host_name)
# Send a bulk of all metrics at once
for (metric, _) in couples:
uri = self.uri + 'render/?width=586&height=308&lineMode=connected&from=' + d + "&until=" + e
if re.search(r'_warn|_crit', metric):
continue
target = "&target=%s%s%s.__HOST__.%s" % (
graphite_pre, host_name, data_source, metric)
uri += target + target + "?????"
v = {}
v['link'] = self.uri
v['img_src'] = uri
v['img_src'] = self._replaceFontSize(v['img_src'],
fontsize[source])
r.append(v)
return r
if t == 'service':
couples = self.get_metric_and_value(elt.perf_data)
# If no values, we can exit now
if len(couples) == 0:
return []
# Remove all non alpha numeric character
desc = self.illegal_char.sub('_', elt.service_description)
host_name = self.illegal_char.sub('_', elt.host.host_name)
# Send a bulk of all metrics at once
for (metric, value) in couples:
uri = self.uri + 'render/?width=586&height=308&lineMode=connected&from=' + d + "&until=" + e
if re.search(r'_warn|_crit', metric):
continue
elif value[1] == '%':
uri += "&yMin=0&yMax=100"
target = "&target=%s%s%s.%s.%s%s" % (graphite_pre, host_name,
data_source, desc, metric,
graphite_post)
uri += target + target + "?????"
v = {}
v['link'] = self.uri
v['img_src'] = uri
v['img_src'] = self._replaceFontSize(v['img_src'],
fontsize.get(source, 10))
r.append(v)
return r
# Oups, bad type?
return []
# REF: doc/broker-modules.png (3)
# We put to external queues broks that was not already send
t0 = time.time()
# We are sending broks as a big list, more efficient than one by one
queues = self.modules_manager.get_external_to_queues()
to_send = [
b for b in self.broks if getattr(b, 'need_send_to_ext', True)
]
for q in queues:
q.put(to_send)
# No more need to send them
for b in to_send:
b.need_send_to_ext = False
logger.debug("Time to send %s broks (%d secs)" %
(len(to_send), time.time() - t0))
# We must had new broks at the end of the list, so we reverse the list
self.broks.reverse()
start = time.time()
while len(self.broks) != 0:
now = time.time()
# Do not 'manage' more than 1s, we must get new broks
# every 1s
if now - start > 1:
break
b = self.broks.pop()
# Ok, we can get the brok, and doing something with it
# REF: doc/broker-modules.png (4-5)
def open(self):
try:
from pymongo import MongoClient
except ImportError:
logger.error(
'[WebUI-MongoDBPreferences] Can not import pymongo.MongoClient'
)
raise
try:
if self.replica_set:
self.con = MongoClient(self.uri,
replicaSet=self.replica_set,
fsync=self.mongodb_fsync)
else:
self.con = MongoClient(self.uri, fsync=self.mongodb_fsync)
logger.info("[WebUI-MongoDBPreferences] connected to mongodb: %s",
self.uri)
self.db = getattr(self.con, self.database)
logger.info(
"[WebUI-MongoDBPreferences] connected to the database: %s",
self.database)
if self.username and self.password:
self.db.authenticate(self.username, self.password)
logger.info(
"[WebUI-MongoDBPreferences] user authenticated: %s",
self.username)
# Check if a document exists in the preferences collection ...
logger.info(
'[WebUI-MongoDBPreferences] searching connection test item in the collection ...'
)
u = self.db.ui_user_preferences.find_one({'_id': 'shinken-test'})
if not u:
# No document ... create a new one!
logger.debug(
'[WebUI-MongoDBPreferences] not found connection test item in the collection'
)
r = self.db.ui_user_preferences.save({
'_id': 'shinken-test',
'last_test': time.time()
})
logger.info(
'[WebUI-MongoDBPreferences] updated connection test item')
else:
# Found document ... update!
logger.debug(
'[WebUI-MongoDBPreferences] found connection test item in the collection'
)
r = self.db.ui_user_preferences.update(
{'_id': 'shinken-test'},
{'$set': {
'last_test': time.time()
}})
logger.info(
'[WebUI-MongoDBPreferences] updated connection test item')
self.is_connected = True
logger.info(
'[WebUI-MongoDBPreferences] database connection established')
except Exception, e:
logger.error("[WebUI-MongoDBPreferences] Exception: %s", str(e))
logger.debug("[WebUI-MongoDBPreferences] Exception type: %s",
type(e))
logger.debug(
"[WebUI-MongoDBPreferences] Back trace of this kill: %s",
traceback.format_exc())
# Depending on exception type, should raise ...
self.is_connected = False
raise
def get_instance(plugin):
logger.debug("[Graphite UI]Get an GRAPHITE UI module for plugin %s" %
plugin.get_name())
instance = Graphite_Webui(plugin)
return instance
def __call__(self, *args):
logger.debug("Calling %s with arguments %s" % (self.f.func_name, args))
return self.f(*args)
except Exception, e:
logger.error("[Ws_arbiter] failed to get the lists: %s" % str(e))
commands_list = []
# We check for auth if it's not anonymously allowed
if app.username != 'anonymous':
basic = parse_auth(request.environ.get('HTTP_AUTHORIZATION', ''))
# Maybe the user not even ask for user/pass. If so, bail out
if not basic:
abort(401, 'Authentication required')
# Maybe he do not give the good credential?
if basic[0] != app.username or basic[1] != app.password:
abort(403, 'Authentication denied')
# Adding commands to the main queue()
logger.debug("[Ws_arbiter] commands = %s" % str(sorted(commands_list)))
for c in sorted(commands_list):
ext = ExternalCommand(c)
app.from_q.put(ext)
# OK here it's ok, it will return a 200 code
# This module will open an HTTP service, where a user can send a command, like a check
# return.
class Ws_arbiter(BaseModule):
def __init__(self, modconf):
BaseModule.__init__(self, modconf)
try:
self.username = getattr(modconf, 'username', 'anonymous')
self.password = getattr(modconf, 'password', '')
try:
p = PerfDatas(s.perf_data)
for m in p:
if m.name and m.value is not None:
logger.debug("[WebUI-cvhost], metric '%s' = %s, uom: %s",
m.name, m.value, m.uom)
if re.search(params['svc_dsk_used'], m.name) and re.match(
params['svc_dsk_uom'], m.uom):
all[m.name] = m.value
logger.debug("[WebUI-cvhost], got '%s' = %s", m.name,
m.value)
except Exception, exp:
logger.warning("[WebUI-cvhost] get_disks, exception: %s", str(exp))
logger.debug("[WebUI-cvhost], get_disks %s", all)
return state, all
def get_memory(h):
all = {}
state = 'UNKNOWN'
s = _findServiceByName(h, params['svc_mem_name'])
if s:
logger.debug("[WebUI-cvhost], found %s", s.get_full_name())
state = s.state
try:
p = PerfDatas(s.perf_data)
for m in p:
logger.error(
"[WebUI-MongoDBPreferences] error during initialization, no database connection!"
)
return None
try:
e = self.db.ui_user_preferences.find_one({'_id': 'shinken-global'})
except Exception, e:
logger.warning("[WebUI-MongoDBPreferences] Exception: %s", str(e))
self.is_connected = False
return None
# Maybe it's a new entry or missing this parameter, bail out
if not e or key not in e:
logger.debug(
"[WebUI-MongoDBPreferences] new parameter of not stored preference: %s",
key)
return None
return e.get(key)
# We will get in the mongodb database the user preference entry, and get the key
# they are asking us
def get_ui_user_preference(self, user, key):
if not self.is_connected:
if not self.open():
logger.error(
"[WebUI-MongoDBPreferences] error during initialization, no database connection!"
)
return None
# You should have received a copy of the GNU Affero General Public License
# along with Shinken. If not, see <http://www.gnu.org/licenses/>.
import os
from config_parser import ConfigParser
from shinken.log import logger
# Get plugin's parameters from configuration file (not useful currently but future ideas ...)
params = {'fake': "fake"}
plugin_name = os.path.splitext(os.path.basename(__file__))[0]
currentdir = os.path.dirname(os.path.realpath(__file__))
configuration_file = "%s/%s" % (currentdir, 'plugin.cfg')
logger.debug("Plugin configuration file: %s", configuration_file)
try:
scp = ConfigParser('#', '=')
params = scp.parse_config(configuration_file)
# mongo_host = params['mongo_host']
params['fake'] = params['fake']
logger.debug("WebUI plugin '%s', configuration loaded.", plugin_name)
# logger.debug("Plugin %s configuration, database: %s (%s)",
# plugin_name, params['mongo_host'], params['mongo_port'])
except Exception as exp:
logger.warning(
"WebUI plugin '%s', configuration file (%s) not available: %s",
plugin_name, configuration_file, str(exp))
def get_page(name):
global params
# user = app.check_user_authentication()
config = 'default'
if '/' in name:
config = name.split('/')[1]
name = name.split('/')[0]
# Find host type if provided in parameters ...
# @mohierf: not yet implemented ...
type = app.request.query.get('type', 'default')
logger.debug("[WebUI-cvhost], get_page for %s (%s)", name,
app.request.query_string)
try:
currentdir = os.path.dirname(os.path.realpath(__file__))
configuration_file = "%s/%s.cfg" % (currentdir, config)
logger.debug("Plugin configuration file: %s", configuration_file)
scp = config_parser('#', '=')
z = params.copy()
z.update(scp.parse_config(configuration_file))
params = z
logger.debug("[WebUI-cvhost] configuration loaded.")
logger.debug("[WebUI-cvhost] configuration, load: %s (%s)",
params['svc_load_name'], params['svc_load_used'])
logger.debug("[WebUI-cvhost] configuration, cpu: %s (%s)",
params['svc_cpu_name'], params['svc_cpu_used'])
logger.debug("[WebUI-cvhost] configuration, disk: %s (%s)",
params['svc_dsk_name'], params['svc_dsk_used'])
logger.debug("[WebUI-cvhost] configuration, memory: %s (%s)",
params['svc_mem_name'], params['svc_mem_used'])
logger.debug("[WebUI-cvhost] configuration, network: %s (%s)",
params['svc_net_name'], params['svc_net_used'])
# logger.debug("[WebUI-cvhost] configuration, printer: %s (%s)", params['svc_prn_name'], params['svc_prn_used'])
except Exception, exp:
logger.warning(
"[WebUI-cvhost] configuration file (%s) not available: %s",
configuration_file, str(exp))
all_perfs = {}
all_states = {}
return {
'app': app,
'config': config,
'all_perfs': all_perfs,
'all_states': all_states
}
