def authentication(self):
try:
if not self.headers.getheader("authorization"):
self.wfile.write('No Authorization Header\n')
return False
else:
request_auth = self.headers.getheader("authorization")
id_pw_list = CONF.rest()['user_password']
try:
request_account = base64.b64decode(str(request_auth).split()[-1])
for id_pw in id_pw_list:
if id_pw.strip() == request_account:
LOG.info('[REST-SERVER] AUTH SUCCESS = %s, from %s', id_pw, self.client_address)
return True
except:
LOG.exception()
self.wfile.write('Request Authentication User ID or Password is Wrong \n')
LOG.info('[REST-SERVER] AUTH FAIL = %s, from %s',
base64.b64decode(str(request_auth).split()[-1]), self.client_address)
return False
except:
LOG.exception()
return False
def stop(self):
try:
pid = self.get_pid()
except IOError:
pid = None
if not pid:
message = "pidfile %s does not exist. Daemon not running?\n"
sys.stderr.write(message % self.pidfile)
return # not an error in a restart
# Try killing the daemon process
try:
LOG.info("--- Daemon STOP ---")
while 1:
for cpid in self.get_child_pid(pid):
os.kill(cpid, SIGTERM)
os.kill(pid, SIGTERM)
time.sleep(0.1)
except OSError, err:
err = str(err)
if err.find("No such process") > 0:
if os.path.exists(self.pidfile):
os.remove(self.pidfile)
else:
print str(err)
print "Stopping Fail ..."
sys.exit(1)
def get_service_list():
service_list = []
try:
url = CONF.xos()['xos_rest_server']
account = CONF.xos()['xos_rest_account']
cmd = 'curl -H "Accept: application/json; indent=4" -u ' + account + ' -X GET ' + url + '/api/core/instances/'
result = Popen(cmd, stdout=PIPE, stderr=PIPE, shell=True)
output, error = result.communicate()
if result.returncode != 0:
LOG.error("Cmd Fail, cause => %s", error)
return ''
instance_array = json.loads(output)
for instance_info in instance_array:
name = instance_info['instance_name']
LOG.info('swarm_instance_name = ' + name)
service_list.append(name)
except:
LOG.exception()
return service_list
def get_disk_usage(username, node_ip, only_value=False):
try:
cmd = 'sudo df -h / | grep -v Filesystem'
cmd_rt = SshCommand.ssh_exec(username, node_ip, cmd)
ratio = float()
if cmd_rt is None:
LOG.info("%s Diksk check Fail", node_ip)
if only_value:
return -1
return {'DISK': 'Command fail'}
else:
if '/' in cmd_rt:
try:
ratio = float(cmd_rt.split()[-2].replace('%', ''))
except:
LOG.exception()
result = {
'DISK': {
'RATIO': float(format(ratio, '.2f')),
'Description': cmd_rt
}
}
LOG.info(" > DISK : %s", str(format(ratio, '.2f')))
if only_value:
return float(format(ratio, '.2f'))
return result
except:
LOG.exception()
return -1
def xos_status_check(conn, db_log, node_name):
xos_status = 'ok'
xos_list = []
fail_reason = []
try:
url = CONF.xos()['xos_rest_server']
account = CONF.xos()['xos_rest_account']
cmd = 'curl -H "Accept: application/json; indent=4" -u ' + account + ' -X GET ' + url + '/api/core/xoses/'
result = Popen(cmd, stdout=PIPE, stderr=PIPE, shell=True)
output, error = result.communicate()
if result.returncode != 0:
LOG.error("Cmd Fail, cause => %s", error)
return 'fail', None
xos_array = json.loads(output)
for xos_info in xos_array:
backend_status = xos_info['backend_status']
LOG.info('xos_status_backend_status = ' + backend_status)
tmp = str(backend_status).split('-')
if tmp[0].strip() == '0':
status = 'ok'
else:
status = 'nok'
xos_json = {
'name': xos_info['name'],
'status': status,
'description': tmp[1].strip()
}
xos_list.append(xos_json)
if status == 'nok':
xos_status = 'nok'
fail_reason.append(xos_json)
try:
sql = 'UPDATE ' + DB.XOS_TBL + \
' SET xos_status = \"' + str(xos_list) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE XOS STATUS INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] XOS STATUS DB Update Fail.')
except:
LOG.exception()
except:
LOG.exception()
xos_status = 'fail'
return xos_status, fail_reason
def auth_pw(self, cli_pw):
id_pw_list = CONF.rest()['user_password']
cli_pw = base64.b64decode(cli_pw)
for id_pw in id_pw_list:
if id_pw.strip() == cli_pw:
LOG.info('[REST-SERVER] AUTH SUCCESS = ' + id_pw)
return True
LOG.info('[REST-SERVER] AUTH FAIL = ' + cli_pw)
return False
def proc_dis_resource(node, param):
res_result = dict()
nodes_info = get_node_list(node, 'nodename, ' + param, DB.RESOURCE_TBL)
LOG.info('*****' + str(nodes_info))
for nodename, value in nodes_info:
if value < 0:
res_result[nodename] = 'FAIL'
else:
res_result[nodename] = value
return res_result
def occur_event(conn, node_name, item, pre_value, cur_value):
time = str(datetime.now())
desc = pre_value + ' -> ' + cur_value
sql = 'UPDATE ' + DB.EVENT_TBL + \
' SET grade = \'' + cur_value + '\'' + ',' + \
' desc = \'' + desc + '\'' + ',' + \
' time = \'' + time + '\'' + \
' WHERE nodename = \'' + node_name + '\' and item = \'' + item + '\''
LOG.info('Update alarm info = ' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
LOG.error('DB Update Fail.')
push_event(node_name, item, cur_value, desc, time)
def get_resource_usage(node_list, param):
res_result = dict()
for node_name, node_ip, username, ping in node_list:
res_result[node_name] = {}
if ping.lower() == 'ok':
LOG.info("GET %s usage for %s", param, node_name)
res_result[node_name] = PARAM_MAP[param](username, node_ip)
else:
LOG.info("Can not get %s Usage... %s Network is NOK", param,
node_name)
res_result[node_name] = 'Net fail'
return res_result
def send_response_traffic_test_old(cond, auth):
trace_result_data = {}
try:
is_success, result = trace.traffic_test_old(cond)
if is_success:
trace_result_data['result'] = 'SUCCESS'
else:
trace_result_data['result'] = 'FAIL'
# trace_result_data['fail_reason'] = 'The source ip does not exist.'
if result != None:
trace_result_data['traffic_test_result'] = result
trace_result_data['transaction_id'] = cond['transaction_id']
try:
LOG.info('%s',
json.dumps(trace_result_data, sort_keys=True, indent=4))
except:
pass
req_body_json = json.dumps(trace_result_data)
try:
url = str(cond['app_rest_url'])
#requests.post(str(url), headers=header, data=req_body_json, timeout=2)
if str(auth).startswith('Basic '):
auth = str(auth).split(' ')[1]
cmd = 'curl -X POST -u \'' + CONF.onos(
)['rest_auth'] + '\' -H \'Content-Type: application/json\' -d \'' + str(
req_body_json) + '\' ' + url
LOG.error('%s', 'curl = ' + cmd)
result = Popen(cmd, stdout=PIPE, stderr=PIPE, shell=True)
result.communicate()
if result.returncode != 0:
# Push noti does not respond
pass
except:
LOG.exception()
pass
except:
LOG.exception()
def exit(self):
try:
pf = file(PIDFILE, 'r')
pid = int(pf.read().strip())
pf.close()
LOG.info("--- Daemon STOP [fail to check rest server] ---")
try:
LOG.info('PID = ' + str(pid))
os.killpg(pid, SIGTERM)
except OSError, err:
err = str(err)
if err.find("No such process") > 0:
if os.path.exists(self.pidfile):
os.remove(self.pidfile)
except:
LOG.exception()
def get_cpu_usage(username, node_ip, only_value=False):
cmd = 'grep \'cpu\ \' /proc/stat'
cmd_rt = SshCommand.ssh_exec(username, node_ip, cmd)
ratio = float()
if cmd_rt is None:
LOG.info("%s CPU check Fail", node_ip)
if only_value:
return -1
return {'CPU': 'Command fail'}
else:
if 'cpu ' in cmd_rt:
LOG.info("cmd_rt: %s", cmd_rt)
try:
f = cmd_rt.split()
ratio = (float(f[1]) + float(f[3])) * 100 / \
(float(f[1]) + float(f[3]) + float(f[4]))
except:
LOG.exception()
result = {
'CPU': {
'RATIO': float(format(ratio, '.2f')),
'Description': cmd_rt
}
}
LOG.info(" CPU check ... %s", result)
if only_value:
return float(format(ratio, '.2f'))
return result
def get_content(self):
if not self.headers.getheader('content-length'):
self.do_HEAD(400)
self.wfile.write(
str({
"result": "FAIL",
"fail_reason": "Bad Request, Content Length is 0\n"
}))
LOG.info('[Data Check] Received No Data from %s',
self.client_address)
return False
else:
try:
receive_data = json.loads(
self.rfile.read(
int(self.headers.getheader("content-length"))))
LOG.info(
'%s', '[Received Data] \n' +
json.dumps(receive_data, sort_keys=True, indent=4))
return receive_data
except:
LOG.exception()
error_reason = 'Json Data Parsing Error\n'
self.do_HEAD(400)
self.wfile.write(
str({
"result": "FAIL",
"fail_reason": error_reason
}))
LOG.info('[Check Content] %s', error_reason)
return False
def get_mem_usage(username, node_ip, only_value=False):
cmd = 'free -t -m | grep Mem'
cmd_rt = SshCommand.ssh_exec(username, node_ip, cmd)
ratio = float()
if cmd_rt is None:
LOG.info("%s Memory check Fail", node_ip)
if only_value:
return -1
return {'MEMORY': 'Command fail'}
else:
if 'Mem' in cmd_rt:
LOG.info("cmd_rt %s", cmd_rt)
try:
f = cmd_rt.split()
ratio = float(f[2]) * 100 / float(f[1])
except:
LOG.exception()
result = {
'MEMORY': {
'RATIO': float(format(ratio, '.2f')),
'Description': cmd_rt
}
}
LOG.info(" Memory check ... %s", result)
if only_value:
return float(format(ratio, '.2f'))
return result
def find_swarm_manager():
hostname = ''
try:
url = CONF.xos()['xos_rest_server']
account = CONF.xos()['xos_rest_account']
cmd = 'curl -H "Accept: application/json; indent=4" -u ' + account + ' -X GET ' + url + '/api/core/controllers/'
result = Popen(cmd, stdout=PIPE, stderr=PIPE, shell=True)
output, error = result.communicate()
if result.returncode != 0:
LOG.error("Cmd Fail, cause => %s", error)
return ''
controller_array = json.loads(output)
for controller_info in controller_array:
auth_url = controller_info['auth_url']
'''
backend_status = controller_info['backend_status']
LOG.info('xos_sync_backend_status = ' + backend_status)
tmp = str(backend_status).split('-')
if tmp[0].strip() == '0':
'''
LOG.info('swarm_manager_auth_url = ' + auth_url)
tmp = str(auth_url).split(':')
hostname = tmp[0]
break
except:
LOG.exception()
return hostname
def run(self):
# DB initiation
DB.db_initiation()
# Start RESTful server
try:
REST_SVR.rest_server_start()
except:
print 'Rest Server failed to start'
LOG.exception()
sys.exit(1)
# Periodic monitoring
if CONF.watchdog()['interval'] == 0:
LOG.info("--- Not running periodic monitoring ---")
while True:
time.sleep(3600)
else:
LOG.info("--- Periodic Monitoring Start ---")
conn = DB.connection()
while True:
try:
watchdog.periodic(conn)
time.sleep(CONF.watchdog()['interval'])
except:
watchdog.push_event('sonawatcher', 'disconnect',
'critical',
'sonawatcher server shutdown',
str(datetime.now()))
conn.close()
LOG.exception()
sys.exit(1)
def tperf_test_run(perf_conditions):
tperf_result = dict()
request_headers = {
'Authorization': CONF.onos()['rest_auth'],
'Accept': 'application/json',
'Content-Type': 'application/json'
}
try:
# 1. creeate instance
LOG.info("[T-perf server/client VM create] --- ")
server_vm, client_vm, client_floatingip = traffic_test.create_instance(
perf_conditions['server'], perf_conditions['client'])
# 2. run performance test
if server_vm and client_vm:
tperf_result = traffic_test.tperf_command_exec(
server_vm.__dict__['addresses'].values()[0][0]['addr'],
client_floatingip.ip, perf_conditions['test_options'])
else:
tperf_result.update({
'result': 'FAIL',
'fail_reason': 'Fail to create instance.'
})
tperf_result.update(
{'transaction_id': perf_conditions['transaction_id']})
LOG.info("[Traffic Performance Test] Return Result = %s",
json.dumps(tperf_result))
# send tperf test result to ONOS
response = requests.post(perf_conditions['app_rest_url'],
data=str(json.dumps(tperf_result)),
headers=request_headers)
LOG.info("[Tperf Result Send] Response = %s %s", response.status_code,
response.reason)
# delete tperf test instance
traffic_test.delete_test_instance(server_vm, client_vm,
client_floatingip)
except:
LOG.exception()
def do_GET(self):
request_sz = int(self.headers["Content-length"])
request_str = self.rfile.read(request_sz)
request_obj = json.loads(request_str)
LOG.info('[REST-SERVER] CLIENT INFO = ' + str(self.client_address))
LOG.info('[REST-SERVER] RECV BODY = \n' +
json.dumps(request_obj, sort_keys=True, indent=4))
if self.headers.getheader('Authorization') is None:
self.do_HEAD(401)
self.wfile.write('no auth header received')
LOG.info('[REST-SERVER] no auth header received')
elif self.auth_pw(self.headers.getheader('Authorization')):
if self.path.startswith('/command'):
if command.exist_command(request_obj):
res_body = command.parse_command(request_obj)
self.do_HEAD(200)
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
else:
self.do_HEAD(404)
self.wfile.write('command not found')
LOG.info('[REST-SERVER] ' + 'command not found')
elif self.path.startswith('/regi'):
try:
self.do_HEAD(200)
url = 'http://' + self.client_address[
0] + ':' + request_obj['port'] + '/' + str(
request_obj['uri'])
res_body = command.regi_url(
url, self.headers.getheader('Authorization'))
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
except:
LOG.exception()
elif self.path.startswith('/unregi'):
try:
self.do_HEAD(200)
url = 'http://' + self.client_address[
0] + ':' + request_obj['port'] + '/' + str(
request_obj['uri'])
res_body = command.unregi_url(url)
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
except:
LOG.exception()
else:
self.do_HEAD(404)
self.wfile.write(self.path + ' not found')
LOG.info('[REST-SERVER] ' + self.path + ' not found')
else:
self.do_HEAD(401)
self.wfile.write(self.headers.getheader('Authorization'))
self.wfile.write('not authenticated')
LOG.info('[REST-SERVER] not authenticated')
def periodic(conn, pre_stat, db_log):
try:
cur_info = {}
#LOG.info('Periodic checking %s', str(CONF.watchdog()['check_system']))
try:
node_list = cmd_proc.get_node_list('all', 'nodename, ip_addr, username, type, sub_type')
if not node_list:
LOG.info("Not Exist Node data ...")
return
except:
LOG.exception()
return
# Read cur alarm status
sql = 'SELECT nodename, item, grade FROM ' + DB.EVENT_TBL
db_log.write_log(sql)
cur_grade = conn.cursor().execute(sql).fetchall()
old_nok_count = 0;
for nodename, item, grade in cur_grade:
if not cur_info.has_key(nodename):
cur_info[nodename] = {}
cur_info[nodename][item] = grade
if grade != 'ok':
old_nok_count += 1
new_nok_count = 0;
for node_name, node_ip, user_name, type, sub_type in node_list:
#LOG.info('------------------------------------ ' + node_name + ' START ------------------------------------')
onos_cluster = 'fail'
onos_device = 'fail'
onos_link = 'fail'
onos_app = 'fail'
# ping check
ping = net_check(node_ip)
ping_reason = []
if ping != 'ok':
reason.append('ping check failed on ' + node_ip)
new_nok_count += 1
ping = alarm_event.process_event(conn, db_log, node_name, type, 'PING', cur_info[node_name]['PING'], ping, ping_reason)
if ping == 'ok':
if type.upper() == 'ONOS':
# check connection
onos_cluster, onos_device, onos_link, onos_app, cluster_reason, device_reason, link_reason, app_reason = chk_onos.onos_check(conn, db_log, node_name, node_ip)
onos_cluster = alarm_event.process_event(conn, db_log, node_name, type, 'ONOS_CLUSTER',
cur_info[node_name]['ONOS_CLUSTER'], onos_cluster, cluster_reason)
onos_device = alarm_event.process_event(conn, db_log, node_name, type, 'ONOS_DEVICE',
cur_info[node_name]['ONOS_DEVICE'], onos_device, device_reason)
onos_link = alarm_event.process_event(conn, db_log, node_name, type, 'ONOS_LINK',
cur_info[node_name]['ONOS_LINK'], onos_link, link_reason)
onos_app = alarm_event.process_event(conn, db_log, node_name, type, 'ONOS_APP',
cur_info[node_name]['ONOS_APP'], onos_app, app_reason)
if onos_cluster != 'ok': new_nok_count += 1
if onos_device != 'ok': new_nok_count += 1
if onos_link != 'ok': new_nok_count += 1
if onos_app != 'ok': new_nok_count += 1
try:
sql = 'UPDATE ' + DB.STATUS_TBL + \
' SET' + \
' PING = \'' + ping + '\',' + \
' ONOS_CLUSTER = \'' + onos_cluster + '\',' + \
' ONOS_DEVICE = \'' + onos_device + '\',' + \
' ONOS_LINK = \'' + onos_link + '\',' + \
' ONOS_APP = \'' + onos_app + '\',' + \
' time = \'' + str(datetime.now()) + '\'' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE TOTAL SYSTEM INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] TOTAL SYSTEM INFO DB Update Fail.')
except:
LOG.exception()
# do not version log on everthing is ok
if old_nok_count > 0:
LOG.info('chk_onos[%s]: ping=%s cluster=%s device=%s link=%s app=%s' %
(node_name, ping, onos_cluster, onos_device, onos_link, onos_app))
if old_nok_count > 0 and new_nok_count == 0:
alarm_event.process_event(conn, db_log, 'ALL', 'SITE', 'STATUS', 'none', 'ok', [])
# send all alarm messages pending
alarm_event.flush_event_alarm();
except:
LOG.exception()
return pre_stat
def do_GET(self):
# health check
if self.path.startswith('/alive-check'):
self.do_HEAD(200)
self.wfile.write('ok\n')
return
if not self.authentication():
self.do_HEAD(401)
return
else:
if not self.headers.getheader('Content-Length'):
self.do_HEAD(400)
self.wfile.write('Bad Request, Content Length is 0\n')
return
else:
request_size = int(self.headers.getheader("Content-Length"))
request_string = self.rfile.read(request_size)
request_obj = json.loads(request_string)
LOG.info('[REST-SERVER] CLIENT INFO = ' + str(self.client_address))
LOG.info('[REST-SERVER] RECV BODY = \n' + json.dumps(request_obj, sort_keys=True, indent=4))
if self.path.startswith('/command'):
try:
if command.exist_command(request_obj):
res_body = command.parse_command(request_obj)
self.do_HEAD(200)
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
else:
self.do_HEAD(404)
self.wfile.write('command not found')
LOG.info('[REST-SERVER] ' + 'command not found')
except:
LOG.exception()
elif self.path.startswith('/regi'):
try:
self.do_HEAD(200)
url = str(request_obj['url'])
res_body = command.regi_url(url, self.headers.getheader('Authorization'))
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
except:
LOG.exception()
elif self.path.startswith('/event_list'):
try:
self.do_HEAD(200)
url = str(request_obj['url'])
res_body = command.get_event_list(url, self.headers.getheader('Authorization'))
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
except:
LOG.exception()
elif self.path.startswith('/unregi'):
try:
self.do_HEAD(200)
url = str(request_obj['url'])
res_body = command.unregi_url(url)
self.wfile.write(json.dumps(res_body))
LOG.info('[REST-SERVER] RES BODY = \n%s',
json.dumps(res_body, sort_keys=True, indent=4))
except:
LOG.exception()
else:
self.do_HEAD(404)
self.wfile.write(self.path + ' not found\n')
LOG.info('[REST-SERVER] ' + self.path + ' not found')
def do_HEAD(self, res_code):
self.send_response(res_code)
self.send_header('Content-type', 'application/json')
self.end_headers()
if res_code != 200:
LOG.info('[REST-SERVER] RESPONSE CODE = ' + str(res_code))
def get_internal_traffic(conn, db_log, node_name, node_ip, user_name, sub_type, rx_count, patch_tx, pre_stat):
try:
status = 'ok'
in_packet = 0
out_packet = 0
reason_list = []
desc = ''
if sub_type == 'COMPUTE':
flow_rt = SshCommand.ssh_exec(user_name, node_ip, 'sudo ovs-ofctl -O OpenFlow13 dump-flows br-int')
inport_cnt = 0
gw_cnt = 0
output_cnt = 0
if flow_rt is not None:
for line in flow_rt.splitlines():
tmp = line.split(',')
if 'in_port' in line:
inport_cnt = inport_cnt + int(tmp[3].split('=')[1])
elif 'output' in line:
output_cnt = output_cnt + int(tmp[3].split('=')[1])
elif 'actions=group' in line:
gw_cnt = gw_cnt + int(tmp[3].split('=')[1])
in_packet = inport_cnt + rx_count
out_packet = gw_cnt + output_cnt
port_json = {'vm_tx': inport_cnt, 'vxlan_rx': rx_count, 'out_gw': gw_cnt, 'output': output_cnt}
else:
port_json = {'vm_tx': -1, 'vxlan_rx': -1, 'out_gw': -1, 'output': -1}
status = 'fail'
else:
port_json = {'vxlan_rx': rx_count, 'patch-integ': patch_tx}
if patch_tx == -1:
status = 'fail'
else:
in_packet = rx_count
out_packet = patch_tx
for_save_in = in_packet
for_save_out = out_packet
if not dict(pre_stat).has_key(node_name + '_internal'):
status = '-'
vxlan_json = {'port_stat_in_out': port_json, 'period': CONF.watchdog()['interval'],
'ratio': 0, 'current_rx': -1, 'current_tx': -1,
'description': desc, 'threshold': CONF.alarm()['internal_traffic_ratio'], 'status': status}
elif status == 'ok':
in_packet = in_packet - int(dict(pre_stat)[node_name + '_internal']['in_packet'])
out_packet = out_packet - int(dict(pre_stat)[node_name + '_internal']['out_packet'])
if in_packet == 0 and out_packet == 0:
ratio = 100
elif in_packet <= 0 or out_packet < 0:
LOG.info('Internal Traffic Ratio Fail.')
ratio = 0
else:
ratio = float(out_packet) * 100 / in_packet
LOG.info('Internal Traffic Ratio = ' + str(ratio))
desc = 'Internal Traffic Ratio = ' + str(ratio) + '(' + str(out_packet) + '/' + str(in_packet) + ')'
if ratio < float(CONF.alarm()['internal_traffic_ratio']):
status = 'nok'
vxlan_json = {'port_stat_in_out': port_json, 'period': CONF.watchdog()['interval'],
'ratio': format(ratio, '.2f'), 'current_rx': in_packet, 'current_tx': out_packet,
'description': desc, 'threshold': CONF.alarm()['internal_traffic_ratio'], 'status': status}
in_out_dic = dict()
in_out_dic['in_packet'] = for_save_in
in_out_dic['out_packet'] = for_save_out
pre_stat[node_name + '_internal'] = in_out_dic
try:
sql = 'UPDATE ' + DB.OPENSTACK_TBL + \
' SET internal_traffic = \"' + str(vxlan_json) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE INTERNAL TRAFFIC INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] INTERNAL TRAFFIC DB Update Fail.')
except:
LOG.exception()
except:
LOG.exception()
status = 'fail'
if not status == 'ok':
reason_list.append(vxlan_json)
return status, pre_stat, reason_list
def xos_sync_check(conn, db_log, node_name):
swarm_sync = 'ok'
sync_list = []
fail_reason = []
try:
url = CONF.xos()['xos_rest_server']
account = CONF.xos()['xos_rest_account']
cmd = 'curl -H "Accept: application/json; indent=4" -u ' + account + ' -X GET ' + url + '/api/core/diags/'
result = Popen(cmd, stdout=PIPE, stderr=PIPE, shell=True)
output, error = result.communicate()
if result.returncode != 0:
LOG.error("Cmd Fail, cause => %s", error)
return 'fail', None
sync_array = json.loads(output)
for xos_info in sync_array:
backend_status = xos_info['backend_status']
LOG.info('xos_sync_backend_status = ' + backend_status)
tmp = str(backend_status).split('-')
if tmp[0].strip() in ['0', '1']:
status = 'ok'
else:
status = 'nok'
# check time
last_time = json.loads(xos_info['backend_register'])['last_run']
cur_time = time.time()
interval = cur_time - last_time
interval = int(interval)
if interval >= 30:
status = 'nok'
xos_json = {
'name': xos_info['name'],
'status': status,
'description': tmp[1].strip(),
'last_run_interval': interval
}
sync_list.append(xos_json)
if status == 'nok':
swarm_sync = 'nok'
fail_reason.append(xos_json)
try:
sql = 'UPDATE ' + DB.XOS_TBL + \
' SET synchronizer = \"' + str(sync_list) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE SYNCHRONIZER INFO -----\n' +
sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] SYNCHRONIZER DB Update Fail.')
except:
LOG.exception()
except:
LOG.exception()
swarm_sync = 'fail'
return swarm_sync, fail_reason
def run(self):
db_log = USER_LOG()
db_log.set_log('db.log',
CONF.base()['log_rotate_time'],
CONF.base()['log_backup_count'])
pre_stat = dict()
# DB initiation
DB.db_initiation(db_log)
# Start RESTful server
try:
REST_SVR.rest_server_start()
except:
print 'Rest Server failed to start'
LOG.exception()
self.exit()
# Periodic monitoring
if CONF.watchdog()['interval'] == 0:
LOG.info("--- Not running periodic monitoring ---")
while True:
time.sleep(3600)
else:
LOG.info("--- Periodic Monitoring Start ---")
history_log.write_log("--- Event History Start ---")
conn = DB.connection()
exitFlag = False
while True:
try:
i = 0
while i < 3:
i = i + 1
# check rest server
try:
url = 'http://' + socket.gethostbyname(
socket.gethostname()) + ':' + str(CONF.rest(
)['rest_server_port']) + '/alive-check'
cmd = 'curl -X GET \"' + url + '\"'
LOG.info('cmd = ' + cmd)
result = Popen(cmd,
stdout=PIPE,
stderr=PIPE,
shell=True)
output, error = result.communicate()
if result.returncode != 0:
LOG.info('REST SERVER CHECK FAIL [' + str(i) +
']')
if i == 3:
LOG.info('fail to check rest server.')
alarm_event.push_event(
'sonawatcher',
'SONAWATCHER_DISCONNECT', 'critical',
'normal',
'sonawatcher server shutdown',
str(datetime.now()))
conn.close()
exitFlag = True
self.exit()
break
else:
break
except:
LOG.exception()
if exitFlag:
break
pre_stat = watchdog.periodic(conn, pre_stat, db_log)
time.sleep(CONF.watchdog()['interval'])
except:
alarm_event.push_event('sonawatcher',
'SONAWATCHER_DISCONNECT',
'critical', 'normal',
'sonawatcher server shutdown',
str(datetime.now()))
conn.close()
LOG.exception()
def get_gw_ratio_gateway(conn, db_log, node_ip, node_name, rx, gw_rx_sum, pre_stat):
status = 'ok'
reason = []
try:
sql = 'SELECT ' + DB.ONOS_TBL + '.nodename, nodelist, ip_addr' + ' FROM ' + DB.ONOS_TBL + \
' INNER JOIN ' + DB.NODE_INFO_TBL + ' ON ' + DB.ONOS_TBL + '.nodename = ' + DB.NODE_INFO_TBL + '.nodename'
nodes_info = conn.cursor().execute(sql).fetchall()
if len(nodes_info) == 0:
LOG.info('Fail to load onos list')
return 'fail', pre_stat, reason
# search data_ip
data_ip = ''
manage_ip = ''
cpt_to_gw_packet = 0
for nodename, nodelist, ip in nodes_info:
if not nodelist == 'none':
for node_info in eval(nodelist):
try:
if dict(node_info)['management_ip'] == node_ip:
manage_ip = ip
data_ip = dict(node_info)['data_ip']
except:
manage_ip = ''
if not manage_ip == '':
break
if not manage_ip == '':
break
if data_ip == '':
LOG.info('Can not find data ip')
return 'fail', pre_stat, reason
group_rt = SshCommand.onos_ssh_exec(manage_ip, 'groups')
if group_rt is not None:
for line in group_rt.splitlines():
if '{tunnelDst=' + data_ip + '}' in line:
tmp = line.split(',')
for col in tmp:
if 'packets=' in col:
cpt_to_gw_packet = cpt_to_gw_packet + int(col.split('=')[1])
if not dict(pre_stat).has_key(node_name + '_GW'):
status = '-'
json_ratio = {'current_rx': '-', 'current_compute_tx': '-', 'current_total': '-',
'ratio': '-',
'period': CONF.watchdog()['interval'], 'status': status, 'packet_loss': False,
'description': ''}
else:
cur_rx = rx - int(dict(pre_stat)[node_name + '_GW']['rx'])
cur_total = gw_rx_sum - int(dict(pre_stat)[node_name + '_GW']['gw_rx_sum'])
cur_packet = cpt_to_gw_packet - int(dict(pre_stat)[node_name + '_GW']['cpt_to_gw_packet'])
if cur_rx == 0 and cur_total == 0:
ratio = 100
elif cur_rx <= 0 or cur_total < 0:
ratio = 0
else:
ratio = float(cur_rx) * 100 / cur_total
desc = 'GW RATIO = ' + str(ratio) + ' (' + str(cur_rx) + ' / ' + str(cur_total) + ')'
loss_flag = False
if cur_rx < cur_packet:
LOG.info('GW Ratio Fail. (Data loss)')
loss_flag = True
LOG.info('GW Ratio = ' + str(ratio))
if ratio < float(CONF.alarm()['gw_ratio']) or cur_rx < cur_packet:
status = 'nok'
json_ratio = {'current_rx': cur_rx, 'current_compute_tx': cur_packet, 'current_total': cur_total, 'ratio': format(ratio, '.2f'),
'period':CONF.watchdog()['interval'], 'status': status, 'packet_loss': loss_flag, 'description': desc}
try:
sql = 'UPDATE ' + DB.OPENSTACK_TBL + \
' SET gw_ratio = \"' + str(json_ratio) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE TRAFFIC GW INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] TRAFFIC GW DB Update Fail.')
except:
LOG.exception()
in_out_dic = dict()
in_out_dic['rx'] = rx
in_out_dic['gw_rx_sum'] = gw_rx_sum
in_out_dic['cpt_to_gw_packet'] = cpt_to_gw_packet
pre_stat[node_name + '_GW'] = in_out_dic
except:
LOG.exception()
status = 'fail'
if not status == 'ok':
reason.append(json_ratio)
return status, pre_stat, reason
def get_gw_ratio_compute(conn, db_log, node_ip, node_name, pre_stat):
status = 'ok'
reason = []
try:
sql = 'SELECT ' + DB.ONOS_TBL + '.nodename, nodelist, ip_addr' + ' FROM ' + DB.ONOS_TBL + \
' INNER JOIN ' + DB.NODE_INFO_TBL + ' ON ' + DB.ONOS_TBL + '.nodename = ' + DB.NODE_INFO_TBL + '.nodename'
nodes_info = conn.cursor().execute(sql).fetchall()
if len(nodes_info) == 0:
LOG.info('Fail to load onos list')
return 'fail', pre_stat, reason
manage_ip = ''
hostname = ''
for nodename, nodelist, ip in nodes_info:
if not nodelist == 'none':
for node_info in eval(nodelist):
try:
if dict(node_info)['management_ip'] == node_ip:
manage_ip = ip
hostname = dict(node_info)['hostname']
except:
manage_ip = ''
if not manage_ip == '':
break
if not manage_ip == '':
break
if hostname == '':
LOG.info('Can not find hostname')
return 'fail', pre_stat, reason
try:
sql = 'SELECT of_id FROM ' + DB.OPENSTACK_TBL + ' WHERE hostname = \'' + str(hostname) + '\''
LOG.info(sql)
node_info = conn.cursor().execute(sql).fetchone()
of_id = node_info[0]
except:
LOG.exception()
LOG.info('Can not find of_id')
return 'fail', pre_stat, reason
group_rt = SshCommand.onos_ssh_exec(manage_ip, 'groups')
total_cnt = 0
gw_list = []
if group_rt is not None:
for line in group_rt.splitlines():
if of_id in line:
tmp = line.split(',')
for col in tmp:
if 'packets=' in col:
total_cnt = total_cnt + int(col.split('=')[1])
gw_list.append(int(col.split('=')[1]))
str_ratio = ''
if not dict(pre_stat).has_key(node_name + '_GW'):
status = '-'
json_ratio = {'ratio': '-', 'status': status, 'period':CONF.watchdog()['interval'], 'status': status}
else:
i = 0
for gw in gw_list:
cur_gw = gw - pre_stat[node_name + '_GW']['gw_list'][i]
cur_total = total_cnt - pre_stat[node_name + '_GW']['gw_total']
LOG.info('cur_gw = ' + str(cur_gw))
LOG.info('cur_total = ' + str(cur_total))
if cur_gw == 0 and cur_total == 0:
ratio = 100/len(gw_list)
elif cur_gw <= 0 or cur_total <= 0:
ratio = 0
else:
ratio = float(cur_gw) * 100 / cur_total
i = i + 1
str_ratio = str_ratio + str(ratio) + ':'
if ratio < float(CONF.alarm()['gw_ratio']):
status = 'nok'
json_ratio = {'ratio': str_ratio.rstrip(':'), 'status': status, 'period':CONF.watchdog()['interval'], 'status': status}
LOG.info('[COMPUTE] ' + 'GW_RATIO = ' + str_ratio.rstrip(':'))
try:
sql = 'UPDATE ' + DB.OPENSTACK_TBL + \
' SET gw_ratio = \"' + str(json_ratio) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE TRAFFIC GW INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] TRAFFIC GW DB Update Fail.')
except:
LOG.exception()
in_out_dic = dict()
in_out_dic['gw_list'] = gw_list
in_out_dic['gw_total'] = total_cnt
pre_stat[node_name + '_GW'] = in_out_dic
except:
LOG.exception()
status = 'fail'
if not status == 'ok':
reason.append(json_ratio)
return status, pre_stat, reason
def vrouter_check(conn, db_log, node_name, user_name, node_ip):
ret_docker = 'ok'
docker_list = []
fail_list = []
onos_id = ''
docker_rt = SshCommand.ssh_exec(user_name, node_ip, 'sudo docker ps')
if docker_rt is not None:
try:
for docker in CONF.openstack()['docker_list']:
for line in docker_rt.splitlines():
if line.startswith('CONTAINER'):
continue
tmp_line = line.split()
if ' ' + docker in line:
if not 'Up' in line:
docker_json = {'name': docker, 'status': 'nok', 'type': 'docker'}
fail_list.append(docker_json)
ret_docker = 'nok'
else:
docker_json = {'name': docker, 'status': 'ok', 'type': 'docker'}
docker_list.append(docker_json)
if 'onos' in tmp_line[1]:
onos_id = tmp_line[0]
except:
LOG.exception()
else:
LOG.error("\'%s\' Vrouter Node Check Error", node_ip)
ret_docker = 'fail'
onos_app_list = []
route_list = []
if not onos_id == '':
try:
# get onos container ip
onos_rt = SshCommand.ssh_exec(user_name, node_ip, 'sudo docker inspect ' + onos_id + ' | grep IPAddress')
if onos_rt is not None:
for line in onos_rt.splitlines():
line = line.strip()
if line.startswith('\"IPAddress'):
tmp = line.split(':')
onos_ip = tmp[1].strip().replace('\"', '').replace(',', '')
break
app_list = SshCommand.ssh_pexpect(user_name, node_ip, onos_ip, 'apps -a -s')
app_active_list = list()
for line in app_list.splitlines():
if line.startswith('fail'):
continue
app_active_list.append(line.split(".")[2].split()[0])
for app in CONF.openstack()['onos_vrouter_app_list']:
if app in app_active_list:
app_json = {'name': app, 'status': 'ok', 'type': 'onos_app'}
else:
app_json = {'name': app, 'status': 'nok', 'type': 'onos_app'}
fail_list.append(app_json)
ret_docker = 'nok'
onos_app_list.append(app_json)
str_route = SshCommand.ssh_pexpect(user_name, node_ip, onos_ip, 'routes')
for line in str_route.splitlines():
line = line.strip()
if (line.startswith('Table') or line.startswith('Network') or line.startswith('Total')):
continue
new_line = " ".join(line.split())
if new_line.startswith('fail'):
continue
tmp = new_line.split(' ')
route_json = {'network': tmp[0], 'next_hop': tmp[1]}
route_list.append(route_json)
except:
LOG.exception()
else:
LOG.info('can not find onos_id.')
ret_docker = 'fail'
try:
sql = 'UPDATE ' + DB.OPENSTACK_TBL + \
' SET docker = \"' + str(docker_list) + '\",' + \
' onosApp = \"' + str(onos_app_list) + '\",' + \
' routingTable = \"' + str(route_list) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE GATEWAY INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] GATEWAY DB Update Fail.')
except:
LOG.exception()
return ret_docker, fail_list
def rest_server_start():
LOG.info("--- REST Server Start --- ")
rest_server_daemon = multiprocess.Process(name='rest_server', target=run)
rest_server_daemon.daemon = True
rest_server_daemon.start()
if pid > 0:
# exit from second parent
sys.exit(0)
except OSError, e:
sys.stderr.write("fork #2 failed: %d (%s)\n" %
(e.errno, e.strerror))
sys.exit(1)
# redirect standard file descriptors
si = file(self.stdin, 'r')
so = file(self.stdout, 'a+')
se = file(self.stderr, 'a+', 0)
pid = str(os.getpid())
LOG.info("--- Daemon START ---")
sys.stderr.write("\nstarted with pid %s\n" % pid)
sys.stderr.flush()
if self.pidfile:
file(self.pidfile, 'w+').write("%s\n" % pid)
atexit.register(self.delpid)
os.dup2(si.fileno(), sys.stdin.fileno())
os.dup2(so.fileno(), sys.stdout.fileno())
os.dup2(se.fileno(), sys.stderr.fileno())
# delete pid file when parent process kill
def delpid(self):
try:
os.remove(self.pidfile)
def get_node_traffic(conn, db_log, node_name, rx_dic, tx_dic, total_rx, total_tx, err_info, pre_stat):
try:
status = 'ok'
reason_list = []
pre_total_rx = total_rx
pre_total_tx = total_tx
# check minimum packet count
sql = 'SELECT data_ip FROM ' + DB.OPENSTACK_TBL + ' WHERE nodename = \'' + node_name + '\''
data_ip = conn.cursor().execute(sql).fetchone()[0]
sql = 'SELECT ip_addr FROM ' + DB.NODE_INFO_TBL + ' WHERE type = \'ONOS\''
nodes_info = conn.cursor().execute(sql).fetchall()
min_rx = 0
if len(nodes_info) == 0:
LOG.info('Fail to load onos list')
status = 'fail'
else:
for ip in nodes_info:
flows_rt = SshCommand.onos_ssh_exec(ip[0], '\"flows --filter \'{tunnelDst=' + data_ip + '}\' --short\"')
if flows_rt is not None:
for line in flows_rt.splitlines():
if 'tunnelDst' in line:
min_rx = min_rx + int(line.split(',')[2].split('=')[1])
break
if not dict(pre_stat).has_key(node_name + '_VXLAN'):
status = '-'
ratio = -1
else:
total_rx = total_rx - int(dict(pre_stat)[node_name + '_VXLAN']['total_rx'])
total_tx = total_tx - int(dict(pre_stat)[node_name + '_VXLAN']['total_tx'])
cur_min = min_rx - int(dict(pre_stat)[node_name + '_VXLAN']['min_rx'])
if total_rx == 0 and total_tx == 0:
ratio = 100
elif total_tx <= 0 or total_tx < 0:
LOG.info('Node Traffic Ratio Fail.')
ratio = 0
else:
ratio = float(total_rx) * 100 / total_tx
LOG.info('Node Traffic Ratio = ' + str(ratio))
port_json = {'rx': rx_dic[node_name], 'minimum_rx': min_rx, 'rx_drop': err_info['rx_drop'], 'rx_errs': err_info['rx_err'],
'tx': tx_dic[node_name], 'tx_drop': err_info['tx_drop'], 'tx_errs': err_info['tx_err']}
description = ''
if not status == '-':
description = 'Ratio of success for all nodes = ' + str(ratio) + ' (' + str(total_rx) + ' / ' + str(total_tx) + ')'
if ratio < float(CONF.alarm()['node_traffic_ratio']):
LOG.info('[NODE TRAFFIC] ratio nok')
status = 'nok'
if total_rx < cur_min:
LOG.info('CUR_MIN_RX = ' + str(cur_min) + ', CUR_RX = ' + str(total_rx) + ', Less than rx minimum.')
status = 'nok'
if err_info['rx_drop'] - int(dict(pre_stat)[node_name + '_VXLAN']['rx_drop']) > 0:
LOG.info('[NODE TRAFFIC] rx_drop nok')
status = 'nok'
if err_info['rx_err'] - int(dict(pre_stat)[node_name + '_VXLAN']['rx_err']) > 0:
LOG.info('[NODE TRAFFIC] rx_err nok')
status = 'nok'
if err_info['tx_drop'] - int(dict(pre_stat)[node_name + '_VXLAN']['tx_drop']) > 0:
LOG.info('[NODE TRAFFIC] tx_drop nok')
status = 'nok'
if err_info['tx_err'] - int(dict(pre_stat)[node_name + '_VXLAN']['tx_err']) > 0:
LOG.info('[NODE TRAFFIC] tx_err nok')
status = 'nok'
in_out_dic = dict()
in_out_dic['total_rx'] = pre_total_rx
in_out_dic['total_tx'] = pre_total_tx
in_out_dic['min_rx'] = min_rx
in_out_dic['rx_drop'] = err_info['rx_drop']
in_out_dic['rx_err'] = err_info['rx_err']
in_out_dic['tx_drop'] = err_info['tx_drop']
in_out_dic['tx_err'] = err_info['tx_err']
pre_stat[node_name + '_VXLAN'] = in_out_dic
except:
LOG.exception()
status = 'fail'
vxlan_json = {'port_stat_vxlan': port_json, 'period': CONF.watchdog()['interval'],
'ratio': format(ratio, '.2f'), 'current_rx': total_rx, 'current_tx': total_tx,
'description': description, 'threshold': CONF.alarm()['node_traffic_ratio'], 'status': status}
try:
sql = 'UPDATE ' + DB.OPENSTACK_TBL + \
' SET vxlan_traffic = \"' + str(vxlan_json) + '\"' + \
' WHERE nodename = \'' + node_name + '\''
db_log.write_log('----- UPDATE VXLAN STAT INFO -----\n' + sql)
if DB.sql_execute(sql, conn) != 'SUCCESS':
db_log.write_log('[FAIL] VXLAN STAT DB Update Fail.')
except:
LOG.exception()
if not status == 'ok':
reason_list.append(vxlan_json)
return status, pre_stat, reason_list