class ConfigDatabaseEventManager(EventManager):
def __init__(self, config, unit_names):
table = 'ObjectConfigDatabaseInfo'
type_info = EventManagerTypeInfo(
object_table=table,
module_type=Module.CONFIG_DATABASE_NODE_MGR,
sandesh_packages=['database.sandesh'])
super(ConfigDatabaseEventManager, self).__init__(
config, type_info, unit_names)
self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(
config.cassandra_repair_logdir, 'config', table,
config.hostip, config.minimum_diskgb,
config.db_port, config.db_jmx_port, config.db_use_ssl,
config.db_user, config.db_password,
self.process_info_manager)
def get_failbits_nodespecific_desc(self, fail_status_bits):
return self.cassandra_mgr.get_failbits_nodespecific_desc(
self, fail_status_bits)
def do_periodic_events(self):
self.cassandra_mgr.database_periodic(self)
# Perform nodetool repair every cassandra_repair_interval hours
if self.tick_count % (60 * self.cassandra_repair_interval) == 0:
self.cassandra_mgr.repair()
super(ConfigDatabaseEventManager, self).do_periodic_events()
class DatabaseEventManager(EventManager):
def __init__(self, config, unit_names):
type_info = EventManagerTypeInfo(object_table="ObjectDatabaseInfo",
module_type=Module.DATABASE_NODE_MGR,
sandesh_packages=['database.sandesh'])
super(DatabaseEventManager, self).__init__(config, type_info,
sandesh_global, unit_names)
# TODO: try to understand is next needed here and use it or remove
#self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(
config.cassandra_repair_logdir, 'analytics',
config.contrail_databases, config.hostip, config.minimum_diskgb,
config.db_port, config.db_jmx_port, config.db_user,
config.db_password, self.process_info_manager)
def get_failbits_nodespecific_desc(self, fail_status_bits):
description = ""
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE:
description += "Disk for analytics db is too low," + \
" cassandra stopped."
if fail_status_bits & self.FAIL_STATUS_SERVER_PORT:
if description != "":
description += " "
description += "Cassandra state detected DOWN."
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE_NA:
description += "Disk space for analytics db not retrievable."
return description
def do_periodic_events(self):
self.cassandra_mgr.database_periodic(self)
super(DatabaseEventManager, self).do_periodic_events()
class ConfigEventManager(EventManager):
def __init__(self, config, unit_names):
type_info = EventManagerTypeInfo(module_type=Module.CONFIG_NODE_MGR,
object_table='ObjectConfigNode',
sandesh_packages=['database.sandesh'])
super(ConfigEventManager, self).__init__(config, type_info,
sandesh_global, unit_names)
self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(
config.cassandra_repair_logdir, 'config',
config.contrail_databases, config.hostip, config.minimum_diskgb,
config.db_port, config.db_jmx_port, config.db_user,
config.db_password, self.process_info_manager)
def get_failbits_nodespecific_desc(self, fail_status_bits):
description = ""
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE:
description += "Disk for config db is too low," + \
" cassandra stopped."
if fail_status_bits & self.FAIL_STATUS_SERVER_PORT:
if description != "":
description += " "
description += "Cassandra state detected DOWN."
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE_NA:
description += "Disk space for config db not retrievable."
return description
def do_periodic_events(self):
self.cassandra_mgr.database_periodic(self)
# Perform nodetool repair every cassandra_repair_interval hours
if self.tick_count % (60 * self.cassandra_repair_interval) == 0:
self.cassandra_mgr.repair()
super(ConfigEventManager, self).do_periodic_events()
class ConfigDatabaseEventManager(EventManager):
def __init__(self, config, unit_names):
table = 'ObjectConfigDatabaseInfo'
type_info = EventManagerTypeInfo(
object_table=table,
module_type=Module.CONFIG_DATABASE_NODE_MGR,
sandesh_packages=['database.sandesh'])
super(ConfigDatabaseEventManager, self).__init__(
config, type_info, unit_names)
self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(
config.cassandra_repair_logdir, 'config', table,
config.hostip, config.minimum_diskgb,
config.db_port, config.db_jmx_port, config.db_use_ssl,
config.db_user, config.db_password,
self.process_info_manager, hostname=config.hostname)
def get_failbits_nodespecific_desc(self, fail_status_bits):
return self.cassandra_mgr.get_failbits_nodespecific_desc(
self, fail_status_bits)
def do_periodic_events(self):
self.cassandra_mgr.database_periodic(self)
# Perform nodetool repair every cassandra_repair_interval hours
if self.tick_count % (60 * self.cassandra_repair_interval) == 0:
self.cassandra_mgr.repair(self)
super(ConfigDatabaseEventManager, self).do_periodic_events()
class AnalyticsDatabaseEventManager(EventManager):
def __init__(self, config, unit_names):
table = 'ObjectDatabaseInfo'
type_info = EventManagerTypeInfo(object_table=table,
module_type=Module.DATABASE_NODE_MGR,
sandesh_packages=['database.sandesh'])
super(AnalyticsDatabaseEventManager,
self).__init__(config, type_info, unit_names)
# TODO: try to understand is next needed here and use it or remove
#self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(config.cassandra_repair_logdir,
'analytics',
table,
config.hostip,
config.minimum_diskgb,
config.db_port,
config.db_jmx_port,
config.db_use_ssl,
config.db_user,
config.db_password,
self.process_info_manager,
hostname=config.hostname)
def get_failbits_nodespecific_desc(self, fail_status_bits):
return self.cassandra_mgr.get_failbits_nodespecific_desc(
self, fail_status_bits)
def do_periodic_events(self):
self.cassandra_mgr.database_periodic(self)
super(AnalyticsDatabaseEventManager, self).do_periodic_events()
class AnalyticsDatabaseEventManager(EventManager):
def __init__(self, config, unit_names):
table = 'ObjectDatabaseInfo'
type_info = EventManagerTypeInfo(
object_table=table,
module_type=Module.DATABASE_NODE_MGR,
sandesh_packages=['database.sandesh'])
super(AnalyticsDatabaseEventManager, self).__init__(
config, type_info, unit_names)
# TODO: try to understand is next needed here and use it or remove
#self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(
config.cassandra_repair_logdir, 'analytics', table,
config.hostip, config.minimum_diskgb,
config.db_port, config.db_jmx_port, config.db_use_ssl,
config.db_user, config.db_password,
self.process_info_manager)
def get_failbits_nodespecific_desc(self, fail_status_bits):
return self.cassandra_mgr.get_failbits_nodespecific_desc(
self, fail_status_bits)
def do_periodic_events(self):
self.cassandra_mgr.database_periodic(self)
super(AnalyticsDatabaseEventManager, self).do_periodic_events()
class DatabaseEventManager(EventManager):
def __init__(self, config, rule_file, unit_names):
if os.path.exists('/tmp/supervisord_database.sock'):
supervisor_serverurl = "unix:///tmp/supervisord_database.sock"
else:
supervisor_serverurl = "unix:///var/run/supervisord_database.sock"
type_info = EventManagerTypeInfo(
package_name='contrail-database-common',
object_table="ObjectDatabaseInfo",
module_type=Module.DATABASE_NODE_MGR,
supervisor_serverurl=supervisor_serverurl,
third_party_processes={
"cassandra": "Dcassandra-pidfile=.*cassandra\.pid",
"zookeeper":
"org.apache.zookeeper.server.quorum.QuorumPeerMain"
},
sandesh_packages=['database.sandesh'])
super(DatabaseEventManager,
self).__init__(config, type_info, rule_file, sandesh_global,
unit_names)
self.hostip = config.hostip
self.db_port = config.db_port
self.minimum_diskgb = config.minimum_diskgb
self.contrail_databases = config.contrail_databases
# TODO: try to understand is next needed for analytics db and use it or remove
#self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(config.cassandra_repair_logdir,
'analyticsDb',
self.contrail_databases,
self.hostip, self.minimum_diskgb,
self.db_port,
self.process_info_manager)
# end __init__
def get_failbits_nodespecific_desc(self, fail_status_bits):
description = ""
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE:
description += "Disk for analytics db is too low," + \
" cassandra stopped."
if fail_status_bits & self.FAIL_STATUS_SERVER_PORT:
if description != "":
description += " "
description += "Cassandra state detected DOWN."
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE_NA:
description += "Disk space for analytics db not retrievable."
return description
def do_periodic_events(self):
if self.cassandra_mgr.can_serve():
self.cassandra_mgr.database_periodic(self)
self.event_tick_60()
# end do_periodic_events
def process(self):
if self.cassandra_mgr.can_serve():
self.cassandra_mgr.process(self)
class ConfigEventManager(EventManager):
def __init__(self, config, rule_file, unit_names):
if os.path.exists('/tmp/supervisord_config.sock'):
supervisor_serverurl = "unix:///tmp/supervisord_config.sock"
else:
supervisor_serverurl = "unix:///var/run/supervisord_config.sock"
type_info = EventManagerTypeInfo(package_name = 'contrail-config',
module_type = Module.CONFIG_NODE_MGR,
object_table = 'ObjectConfigNode',
supervisor_serverurl = supervisor_serverurl,
third_party_processes = {
"cassandra" : "Dcassandra-pidfile=.*cassandra\.pid",
"zookeeper" : "org.apache.zookeeper.server.quorum.QuorumPeerMain"
},
sandesh_packages = ['database.sandesh'])
super(ConfigEventManager, self).__init__(config, type_info, rule_file,
sandesh_global, unit_names)
self.hostip = config.hostip
self.db_port = config.db_port
self.minimum_diskgb = config.minimum_diskgb
self.contrail_databases = config.contrail_databases
self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_mgr = CassandraManager(config.cassandra_repair_logdir,
'configDb', self.contrail_databases,
self.hostip, self.minimum_diskgb,
self.db_port, self.process_info_manager)
if self.cassandra_mgr.can_serve():
self.add_unit_name('contrail-database.service')
# end __init__
def get_failbits_nodespecific_desc(self, fail_status_bits):
description = ""
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE:
description += "Disk for config db is too low," + \
" cassandra stopped."
if fail_status_bits & self.FAIL_STATUS_SERVER_PORT:
if description != "":
description += " "
description += "Cassandra state detected DOWN."
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE_NA:
description += "Disk space for config db not retrievable."
return description
def do_periodic_events(self):
if self.cassandra_mgr.can_serve():
self.cassandra_mgr.database_periodic(self)
# Perform nodetool repair every cassandra_repair_interval hours
if self.tick_count % (60 * self.cassandra_repair_interval) == 0:
self.cassandra_mgr.repair()
self.event_tick_60()
# end do_periodic_events
def process(self):
if self.cassandra_mgr.can_serve():
self.cassandra_mgr.process(self)
class ConfigEventManager(EventManager):
def __init__(self, config, rule_file, unit_names):
if os.path.exists('/tmp/supervisord_config.sock'):
supervisor_serverurl = "unix:///tmp/supervisord_config.sock"
else:
supervisor_serverurl = "unix:///var/run/supervisord_config.sock"
self.db = package_installed('contrail-openstack-database')
self.config_db = package_installed('contrail-database-common')
if not self.db and self.config_db:
unit_names.append('contrail-database.service')
type_info = EventManagerTypeInfo(package_name = 'contrail-config',
module_type = Module.CONFIG_NODE_MGR,
object_table = 'ObjectConfigNode',
supervisor_serverurl = supervisor_serverurl,
third_party_processes = {
"cassandra" : "Dcassandra-pidfile=.*cassandra\.pid",
"zookeeper" : "org.apache.zookeeper.server.quorum.QuorumPeerMain"
},
sandesh_packages = ['database.sandesh'],
unit_names = unit_names)
super(ConfigEventManager, self).__init__(config, type_info, rule_file,
sandesh_global)
self.hostip = config.hostip
self.db_port = config.db_port
self.minimum_diskgb = config.minimum_diskgb
self.contrail_databases = config.contrail_databases
self.cassandra_repair_interval = config.cassandra_repair_interval
self.cassandra_repair_logdir = config.cassandra_repair_logdir
self.cassandra_mgr = CassandraManager(self.cassandra_repair_logdir,
'configDb', self.contrail_databases,
self.hostip, self.minimum_diskgb,
self.db_port)
# end __init__
def get_failbits_nodespecific_desc(self, fail_status_bits):
description = ""
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE:
description += "Disk for config db is too low," + \
" cassandra stopped."
if fail_status_bits & self.FAIL_STATUS_SERVER_PORT:
if description != "":
description += " "
description += "Cassandra state detected DOWN."
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE_NA:
description += "Disk space for config db not retrievable."
return description
def do_periodic_events(self):
if not self.db and self.config_db:
self.cassandra_mgr.database_periodic(self)
# Perform nodetool repair every cassandra_repair_interval hours
if self.tick_count % (60 * self.cassandra_repair_interval) == 0:
self.cassandra_mgr.repair()
self.event_tick_60()
# end do_periodic_events
def process(self):
if not self.db and self.config_db:
self.cassandra_mgr.process(self)
class DatabaseEventManager(EventManager):
def __init__(self, rule_file, discovery_server,
discovery_port, collector_addr,
hostip, db_port, minimum_diskgb, contrail_databases,
cassandra_repair_interval,
cassandra_repair_logdir, **dss_kwargs):
self.node_type = "contrail-database"
self.uve_node_type = UVENodeTypeNames[NodeType.DATABASE]
self.table = "ObjectDatabaseInfo"
self.module = Module.DATABASE_NODE_MGR
self.module_id = ModuleNames[self.module]
self.hostip = hostip
self.db_port = db_port
self.minimum_diskgb = minimum_diskgb
self.contrail_databases = contrail_databases
self.cassandra_repair_interval = cassandra_repair_interval
self.cassandra_repair_logdir = cassandra_repair_logdir
self.cassandra_mgr = CassandraManager(self.cassandra_repair_logdir,
'analyticsDb', self.table,
self.contrail_databases,
self.hostip, self.minimum_diskgb,
self.db_port)
if os.path.exists('/tmp/supervisord_database.sock'):
self.supervisor_serverurl = "unix:///tmp/supervisord_database.sock"
else:
self.supervisor_serverurl = "unix:///var/run/supervisord_database.sock"
self.add_current_process()
node_type = Module2NodeType[self.module]
node_type_name = NodeTypeNames[node_type]
self.sandesh_global = sandesh_global
EventManager.__init__(
self, rule_file, discovery_server,
discovery_port, collector_addr, sandesh_global, **dss_kwargs)
self.sandesh_global = sandesh_global
if self.rule_file is '':
self.rule_file = "/etc/contrail/" + \
"supervisord_database_files/contrail-database.rules"
json_file = open(self.rule_file)
self.rules_data = json.load(json_file)
_disc = self.get_discovery_client()
sandesh_global.init_generator(
self.module_id, socket.gethostname(), node_type_name,
self.instance_id, self.collector_addr, self.module_id, 8103,
['database.sandesh', 'nodemgr.common.sandesh'], _disc)
sandesh_global.set_logging_params(enable_local_log=True)
ConnectionState.init(sandesh_global, socket.gethostname(), self.module_id,
self.instance_id,
staticmethod(ConnectionState.get_conn_state_cb),
NodeStatusUVE, NodeStatus, self.table,
self.get_process_state_cb)
self.send_init_info()
self.third_party_process_dict = {}
self.third_party_process_dict["cassandra"] = "Dcassandra-pidfile=.*cassandra\.pid"
self.third_party_process_dict["zookeeper"] = "org.apache.zookeeper.server.quorum.QuorumPeerMain"
# end __init__
def _get_cassandra_config_option(self, config):
(linux_dist, x, y) = platform.linux_distribution()
if (linux_dist in ['Ubuntu', 'debian']):
yamlstream = open("/etc/cassandra/cassandra.yaml", 'r')
else:
yamlstream = open("/etc/cassandra/conf/cassandra.yaml", 'r')
cfg = yaml.safe_load(yamlstream)
yamlstream.close()
return cfg[config]
def msg_log(self, msg, level):
self.sandesh_global.logger().log(SandeshLogger.get_py_logger_level(
level), msg)
def send_process_state_db(self, group_names):
self.send_process_state_db_base(
group_names, ProcessInfo)
def send_nodemgr_process_status(self):
self.send_nodemgr_process_status_base(
ProcessStateNames, ProcessState, ProcessStatus)
def get_node_third_party_process_dict(self):
return self.third_party_process_dict
def get_process_state(self, fail_status_bits):
return self.get_process_state_base(
fail_status_bits, ProcessStateNames, ProcessState)
def get_failbits_nodespecific_desc(self, fail_status_bits):
description = ""
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE:
description += "Disk for analytics db is too low," + \
" cassandra stopped."
if fail_status_bits & self.FAIL_STATUS_SERVER_PORT:
if description != "":
description += " "
description += "Cassandra state detected DOWN."
if fail_status_bits & self.FAIL_STATUS_DISK_SPACE_NA:
description += "Disk space for analytics db not retrievable."
return description
def runforever(self, test=False):
self.prev_current_time = int(time.time())
# Initialize tpstat structures
self.cassandra_status_old = CassandraStatusData()
self.cassandra_status_old.cassandra_compaction_task = CassandraCompactionTask()
self.cassandra_status_old.thread_pool_stats = []
while 1:
# we explicitly use self.stdin, self.stdout, and self.stderr
# instead of sys.* so we can unit test this code
headers, payload = self.listener_nodemgr.wait(
self.stdin, self.stdout)
# self.stderr.write("headers:\n" + str(headers) + '\n')
# self.stderr.write("payload:\n" + str(payload) + '\n')
pheaders, pdata = childutils.eventdata(payload + '\n')
# self.stderr.write("pheaders:\n" + str(pheaders)+'\n')
# self.stderr.write("pdata:\n" + str(pdata))
# check for process state change events
if headers['eventname'].startswith("PROCESS_STATE"):
self.event_process_state(pheaders, headers)
# check for flag value change events
if headers['eventname'].startswith("PROCESS_COMMUNICATION"):
self.event_process_communication(pdata)
# do periodic events
if headers['eventname'].startswith("TICK_60"):
self.cassandra_mgr.database_periodic(self)
self.event_tick_60()
self.listener_nodemgr.ok(self.stdout)
def get_package_name(self):
return self.node_type
def process(self):
self.cassandra_mgr.process(self)