def _build_configs_and_plugin():
c1 = p.Config('n-1', 'at-1', 'cluster')
c2 = p.Config('n-2', 'at-2', 'cluster')
c3 = p.Config('n-3', 'at-1', 'node')
class TestPlugin(TestEmptyPlugin):
def get_configs(self, hadoop_version):
return [c1, c2, c3]
return c1, c2, c3, TestPlugin()
def _initialize(self, config):
for configuration in self.config['configurations']:
for service_property in configuration['properties']:
config = p.Config(service_property['name'],
self._get_target(
service_property['applicable_target']),
service_property['scope'],
config_type=
service_property['config_type'],
default_value=service_property
['default_value'],
is_optional=service_property[
'is_optional'],
description=service_property[
'description'])
setattr(config, 'tag', configuration['tag'].rsplit(".", 1)[0])
self.config_items.append(config)
#TODO(jspeidel): an assumption is made that property names
# are unique across configuration sections which is dangerous
property_name = service_property['name']
# if property already exists, throw an exception
if property_name in self.config_mapper:
# internal error
# ambari-config-resource contains duplicates
raise InvalidDataException(
'Internal Error. Duplicate property '
'name detected: %s' % property_name)
self.config_mapper[service_property['name']] = \
self._get_target(
service_property['applicable_target'])
def _initialise_configs():
configs = []
for service, config_lists in XML_CONFS.iteritems():
for config_list in config_lists:
for config in config_list:
if config['name'] not in HIDDEN_CONFS:
cfg = p.Config(config['name'],
service,
"cluster",
is_optional=True,
config_type="string",
default_value=str(config['value']),
description=config['description'])
if config.get('type'):
cfg.config_type = CFG_TYPE[config['type']]
if cfg.config_type == 'bool':
cfg.default_value = cfg.default_value == 'true'
if cfg.config_type == 'int':
if cfg.default_value:
cfg.default_value = int(cfg.default_value)
else:
cfg.config_type = 'string'
if config['name'] in PRIORITY_1_CONFS:
cfg.priority = 1
configs.append(cfg)
configs.append(IDH_TARBALL_URL)
configs.append(IDH_REPO_URL)
configs.append(OS_REPO_URL)
configs.append(OOZIE_EXT22_URL)
configs.append(ENABLE_SWIFT)
return configs
def _initialise_configs():
configs = []
for service, config_lists in XML_CONFS.iteritems():
for config_list in config_lists:
for config in config_list:
if config['name'] not in HIDDEN_CONFS:
cfg = p.Config(config['name'],
service,
"node",
is_optional=True,
config_type="string",
default_value=str(config['value']),
description=config['description'])
if cfg.default_value in ["true", "false"]:
cfg.config_type = "bool"
cfg.default_value = (cfg.default_value == 'true')
elif types.is_int(cfg.default_value):
cfg.config_type = "int"
cfg.default_value = int(cfg.default_value)
if config['name'] in CLUSTER_WIDE_CONFS:
cfg.scope = 'cluster'
if config['name'] in PRIORITY_1_CONFS:
cfg.priority = 1
configs.append(cfg)
for service, config_items in ENV_CONFS.iteritems():
for name, param_format_str in config_items.iteritems():
configs.append(
p.Config(name,
service,
"node",
default_value=1024,
priority=1,
config_type="int"))
configs.append(ENABLE_SWIFT)
configs.append(ENABLE_MYSQL)
if CONF.enable_data_locality:
configs.append(ENABLE_DATA_LOCALITY)
return configs
'Job Tracker Heap Size': 'HADOOP_JOBTRACKER_OPTS=\\"-Xmx%sm\\"',
'Task Tracker Heap Size': 'HADOOP_TASKTRACKER_OPTS=\\"-Xmx%sm\\"'
},
"HDFS": {
'Name Node Heap Size': 'HADOOP_NAMENODE_OPTS=\\"-Xmx%sm\\"',
'Data Node Heap Size': 'HADOOP_DATANODE_OPTS=\\"-Xmx%sm\\"'
},
"JobFlow": {
'Oozie Heap Size': 'CATALINA_OPTS=\\"-Xmx%sm\\"'
}
}
ENABLE_SWIFT = p.Config('Enable Swift',
'general',
'cluster',
config_type="bool",
priority=1,
default_value=True,
is_optional=True)
HIDDEN_CONFS = [
'fs.default.name', 'dfs.name.dir', 'dfs.data.dir', 'mapred.job.tracker',
'mapred.system.dir', 'mapred.local.dir', 'hadoop.proxyuser.hadoop.hosts',
'hadoop.proxyuser.hadoop.groups'
]
CLUSTER_WIDE_CONFS = [
'dfs.block.size', 'dfs.permissions', 'dfs.replication',
'dfs.replication.min', 'dfs.replication.max', 'io.file.buffer.size',
'mapreduce.job.counters.max', 'mapred.output.compress',
'io.compression.codecs', 'mapred.output.compression.codec',
'Job Tracker Heap Size': 'HADOOP_JOBTRACKER_OPTS=\\"-Xmx%sm\\"',
'Task Tracker Heap Size': 'HADOOP_TASKTRACKER_OPTS=\\"-Xmx%sm\\"'
},
"HDFS": {
'Name Node Heap Size': 'HADOOP_NAMENODE_OPTS=\\"-Xmx%sm\\"',
'Data Node Heap Size': 'HADOOP_DATANODE_OPTS=\\"-Xmx%sm\\"'
},
"JobFlow": {
'Oozie Heap Size': 'CATALINA_OPTS=\\"-Xmx%sm\\"'
}
}
ENABLE_SWIFT = p.Config('Enable Swift',
'general',
'cluster',
config_type="bool",
priority=1,
default_value=True,
is_optional=True)
ENABLE_DATA_LOCALITY = p.Config('Enable Data Locality',
'general',
'cluster',
config_type="bool",
priority=1,
default_value=True,
is_optional=True)
ENABLE_MYSQL = p.Config('Enable MySQL',
'general',
'cluster',
'plugins/intel/v2_5_1/resources/mapred-default.xml')
OOZIE_DEFAULT = x.load_hadoop_xml_defaults(
'plugins/intel/v2_5_1/resources/oozie-default.xml')
XML_CONFS = {
"Hadoop": [CORE_DEFAULT],
"HDFS": [HDFS_DEFAULT],
"MapReduce": [MAPRED_DEFAULT],
"JobFlow": [OOZIE_DEFAULT]
}
IDH_TARBALL_URL = p.Config('IDH tarball URL',
'general',
'cluster',
priority=1,
default_value='http://repo1.intelhadoop.com:3424/'
'setup/setup-intelhadoop-'
'2.5.1-en-evaluation.RHEL.tar.gz')
OS_REPO_URL = p.Config('OS repository URL',
'general',
'cluster',
priority=1,
is_optional=True,
default_value='http://mirror.centos.org/'
'centos-6/6/os/x86_64')
IDH_REPO_URL = p.Config('IDH repository URL',
'general',
'cluster',
