deb-sahara/sahara/plugins/hdp/ambariplugin.py
Vitaly Gridnev fbcf8d979e Use ThreadGroup instead of separate threads
ThreadGroup can be used for provision_ambari execution. In that case
we will wait until successful execution of provision_ambari for all
servers.

Change-Id: I60913824e15add981857caef7d5bc94cae5f9512
Closes-bug: 1448073
2015-04-27 13:18:37 +03:00

451 lines
18 KiB
Python

# Copyright (c) 2013 Hortonworks, Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from oslo_config import cfg
from oslo_log import log as logging
from sahara import conductor as c
from sahara import context
from sahara.i18n import _
from sahara.i18n import _LI
from sahara.plugins import exceptions as ex
from sahara.plugins.hdp import hadoopserver as h
from sahara.plugins.hdp import saharautils as utils
from sahara.plugins.hdp.versions import versionhandlerfactory as vhf
from sahara.plugins import provisioning as p
from sahara.topology import topology_helper as th
from sahara.utils import cluster_progress_ops as cpo
from sahara.utils import general as g
conductor = c.API
LOG = logging.getLogger(__name__)
CONF = cfg.CONF
class AmbariPlugin(p.ProvisioningPluginBase):
def __init__(self):
self.cluster_ambari_mapping = {}
self.version_factory = vhf.VersionHandlerFactory.get_instance()
def create_cluster(self, cluster):
version = cluster.hadoop_version
handler = self.version_factory.get_version_handler(version)
cluster_spec = handler.get_cluster_spec(
cluster, self._map_to_user_inputs(
version, cluster.cluster_configs))
hosts = self._get_servers(cluster)
ambari_info = self.get_ambari_info(cluster_spec)
self.cluster_ambari_mapping[cluster.name] = ambari_info
rpm = self._get_rpm_uri(cluster_spec)
servers = []
for host in hosts:
host_role = utils.get_host_role(host)
servers.append(
h.HadoopServer(host, cluster_spec.node_groups[host_role],
ambari_rpm=rpm))
self._provision_cluster(
cluster.name, cluster_spec, ambari_info, servers,
cluster.hadoop_version)
# add the topology data file and script if rack awareness is
# enabled
self._configure_topology_for_cluster(cluster, servers)
LOG.info(_LI("Install of Hadoop stack successful."))
# add service urls
self._set_cluster_info(cluster, cluster_spec)
# check if HDFS HA is enabled; set it up if so
if cluster_spec.is_hdfs_ha_enabled(cluster):
cluster = g.change_cluster_status(cluster, "Configuring HA")
self.configure_hdfs_ha(cluster)
@cpo.event_wrapper(
True, step=_("Add configurations to cluster"), param=('cluster', 1))
def configure_hdfs_ha(self, cluster):
version = cluster.hadoop_version
handler = self.version_factory.get_version_handler(version)
cluster_spec = handler.get_cluster_spec(
cluster, self._map_to_user_inputs(
version, cluster.cluster_configs))
hosts = self._get_servers(cluster)
ambari_info = self.get_ambari_info(cluster_spec)
self.cluster_ambari_mapping[cluster.name] = ambari_info
rpm = self._get_rpm_uri(cluster_spec)
servers = []
for host in hosts:
host_role = utils.get_host_role(host)
servers.append(
h.HadoopServer(host, cluster_spec.node_groups[host_role],
ambari_rpm=rpm))
ambari_client = handler.get_ambari_client()
ambari_client.setup_hdfs_ha(cluster_spec, servers, ambari_info,
cluster.name)
def _get_servers(self, cluster):
servers = []
if hasattr(cluster, 'node_groups') and cluster.node_groups is not None:
# code for a cluster object
for node_group in cluster.node_groups:
servers += node_group.instances
else:
# cluster is actually a cloud context
servers = cluster.instances
return servers
def get_node_processes(self, hadoop_version):
node_processes = {}
version_handler = (
self.version_factory.get_version_handler(hadoop_version))
default_config = version_handler.get_default_cluster_configuration()
for service in default_config.services:
components = []
for component in service.components:
if service.is_user_template_component(component):
components.append(component.name)
node_processes[service.name] = components
return node_processes
def convert(self, config, plugin_name, version, template_name,
cluster_template_create):
handler = self.version_factory.get_version_handler(version)
normalized_config = handler.get_cluster_spec(
None, None, cluster_template=config).normalize()
node_groups = []
for ng in normalized_config.node_groups:
node_group = {
"name": ng.name,
"flavor_id": ng.flavor,
"node_processes": ng.node_processes,
"count": ng.count
}
node_groups.append(node_group)
cluster_configs = dict()
config_resource = handler.get_config_items()
for entry in normalized_config.cluster_configs:
user_input = next((ui for ui in config_resource
if entry.config.name == ui.name), None)
if user_input is not None:
ci = entry.config
# get the associated service dictionary
target = entry.config.applicable_target
service_dict = cluster_configs.get(target, {})
service_dict[ci.name] = entry.value
cluster_configs[target] = service_dict
else:
LOG.debug('Template based input "{entry_name}" is being'
' filtered out as it is not considered a user input'
.format(entry_name=entry.config.name))
ctx = context.ctx()
return cluster_template_create(ctx,
{"name": template_name,
"plugin_name": plugin_name,
"hadoop_version": version,
"node_groups": node_groups,
"cluster_configs": cluster_configs})
def update_infra(self, cluster):
pass
def convert_props_to_template(self, props):
raise NotImplementedError('not yet supported')
def _provision_cluster(self, name, cluster_spec, ambari_info,
servers, version):
# TODO(jspeidel): encapsulate in another class
if servers:
cpo.add_provisioning_step(
servers[0].cluster_id,
_("Provision cluster via Ambari"), len(servers))
with context.ThreadGroup() as tg:
for server in servers:
with context.set_current_instance_id(
server.instance['instance_id']):
tg.spawn(
"hdp-provision-instance-%s" %
server.instance.hostname(),
server.provision_ambari, ambari_info, cluster_spec)
handler = self.version_factory.get_version_handler(version)
ambari_client = handler.get_ambari_client()
ambari_client.wait_for_host_registrations(len(servers), ambari_info)
self._set_ambari_credentials(cluster_spec, ambari_info, version)
ambari_client.provision_cluster(
cluster_spec, servers, ambari_info, name)
LOG.info(_LI('Cluster provisioned via Ambari Server: {server_ip}')
.format(server_ip=ambari_info.get_address()))
# TODO(jspeidel): invoke during scale cluster. Will need to handle dups
def _set_cluster_info(self, cluster, cluster_spec):
info = {}
for service in cluster_spec.services:
if service.deployed:
service.register_service_urls(cluster_spec, info)
conductor.cluster_update(context.ctx(), cluster, {'info': info})
def _set_ambari_credentials(self, cluster_spec, ambari_info, version):
services = cluster_spec.services
ambari_client = (self.version_factory.get_version_handler(version).
get_ambari_client())
for service in services:
if service.name == 'AMBARI':
is_admin_provided = False
admin_user = ambari_info.user
admin_password = ambari_info.password
for user in service.users:
if user.name == 'admin':
ambari_client.update_ambari_admin_user(
user.password, ambari_info)
is_admin_provided = True
ambari_info.user = 'admin'
ambari_info.password = user.password
else:
ambari_client.add_ambari_user(user, ambari_info)
if 'admin' in user.groups:
admin_user = user.name
admin_password = user.password
if not is_admin_provided:
if admin_user is None:
raise ex.HadoopProvisionError(_("An Ambari user in the"
" admin group must be "
"configured."))
ambari_info.user = admin_user
ambari_info.password = admin_password
ambari_client.delete_ambari_user('admin', ambari_info)
break
def _update_ambari_info_credentials(self, cluster_spec, ambari_info):
services = cluster_spec.services
ambari_service = next((service for service in services if
service.name == 'AMBARI'), None)
if ambari_service is not None:
admin_user = next((user for user in ambari_service.users
if 'admin' in user.groups), None)
if admin_user is not None:
ambari_info.user = admin_user.name
ambari_info.password = admin_user.password
LOG.info(_LI('Using "{username}" as admin user for scaling of cluster')
.format(username=ambari_info.user))
# PLUGIN SPI METHODS:
def get_versions(self):
return self.version_factory.get_versions()
def configure_cluster(self, cluster):
self.create_cluster(cluster)
def get_configs(self, hadoop_version):
handler = self.version_factory.get_version_handler(hadoop_version)
return handler.get_config_items()
# cluster name argument supports the non-sahara cluster creation mode
def start_cluster(self, cluster):
client = self.version_factory.get_version_handler(
cluster.hadoop_version).get_ambari_client()
handler = self.version_factory.get_version_handler(
cluster.hadoop_version)
cluster_spec = handler.get_cluster_spec(
cluster, self._map_to_user_inputs(
cluster.hadoop_version, cluster.cluster_configs))
try:
client.start_services(cluster.name, cluster_spec,
self.cluster_ambari_mapping[cluster.name])
finally:
client.cleanup(self.cluster_ambari_mapping[cluster.name])
def get_title(self):
return 'Hortonworks Data Platform'
def get_description(self):
return _('The Hortonworks Sahara plugin automates the deployment '
'of the Hortonworks Data Platform (HDP) on OpenStack.')
def validate(self, cluster):
# creating operational config results in validation
handler = self.version_factory.get_version_handler(
cluster.hadoop_version)
handler.get_cluster_spec(cluster, self._map_to_user_inputs(
cluster.hadoop_version, cluster.cluster_configs))
def scale_cluster(self, cluster, instances):
handler = self.version_factory.get_version_handler(
cluster.hadoop_version)
ambari_client = handler.get_ambari_client()
cluster_spec = handler.get_cluster_spec(
cluster, self._map_to_user_inputs(
cluster.hadoop_version, cluster.cluster_configs))
rpm = self._get_rpm_uri(cluster_spec)
servers = []
for instance in instances:
host_role = utils.get_host_role(instance)
servers.append(h.HadoopServer(instance,
cluster_spec.node_groups
[host_role],
ambari_rpm=rpm))
ambari_info = self.get_ambari_info(cluster_spec)
self._update_ambari_info_credentials(cluster_spec, ambari_info)
cpo.add_provisioning_step(
cluster.id, _("Provision cluster via Ambari"), len(servers))
with context.ThreadGroup() as tg:
for server in servers:
with context.set_current_instance_id(
server.instance['instance_id']):
tg.spawn('Ambari provisioning thread',
server.provision_ambari,
ambari_info, cluster_spec)
ambari_client.configure_scaled_cluster_instances(
cluster.name, cluster_spec, self._get_num_hosts(cluster),
ambari_info)
self._configure_topology_for_cluster(cluster, servers)
ambari_client.start_scaled_cluster_instances(cluster.name,
cluster_spec, servers,
ambari_info)
ambari_client.cleanup(ambari_info)
def get_edp_engine(self, cluster, job_type):
version_handler = (
self.version_factory.get_version_handler(cluster.hadoop_version))
return version_handler.get_edp_engine(cluster, job_type)
def get_edp_job_types(self, versions=[]):
res = {}
for vers in self.version_factory.get_versions():
if not versions or vers in versions:
vh = self.version_factory.get_version_handler(vers)
res[vers] = vh.get_edp_job_types()
return res
def get_edp_config_hints(self, job_type, version):
version_handler = (
self.version_factory.get_version_handler(version))
return version_handler.get_edp_config_hints(job_type)
def decommission_nodes(self, cluster, instances):
LOG.info(_LI('AmbariPlugin: decommission_nodes called for '
'HDP version = {version}')
.format(version=cluster.hadoop_version))
handler = self.version_factory.get_version_handler(
cluster.hadoop_version)
ambari_client = handler.get_ambari_client()
cluster_spec = handler.get_cluster_spec(
cluster, self._map_to_user_inputs(
cluster.hadoop_version, cluster.cluster_configs))
ambari_info = self.get_ambari_info(cluster_spec)
ambari_client.decommission_cluster_instances(cluster, cluster_spec,
instances,
ambari_info)
def validate_scaling(self, cluster, existing, additional):
handler = self.version_factory.get_version_handler(
cluster.hadoop_version)
# results in validation
handler.get_cluster_spec(
cluster, [], dict(existing.items() + additional.items()))
def _get_num_hosts(self, cluster):
count = 0
for node_group in cluster.node_groups:
count += node_group.count
return count
def _get_host_list(self, servers):
host_list = [server.instance.fqdn().lower() for server in servers]
return ",".join(host_list)
def _get_rpm_uri(self, cluster_spec):
ambari_config = cluster_spec.configurations['ambari']
return ambari_config.get('rpm', None)
def get_ambari_info(self, cluster_spec):
ambari_host = cluster_spec.determine_component_hosts(
'AMBARI_SERVER').pop()
port = cluster_spec.configurations['ambari'].get(
'server.port', '8080')
return AmbariInfo(ambari_host, port, 'admin', 'admin')
def _configure_topology_for_cluster(self, cluster, servers):
if CONF.enable_data_locality:
cpo.add_provisioning_step(
cluster.id, _("Enable data locality for cluster"),
len(servers))
topology_data = th.generate_topology_map(
cluster, CONF.enable_hypervisor_awareness)
topology_str = "\n".join(
[k + " " + v for k, v in topology_data.items()]) + "\n"
for server in servers:
server.configure_topology(topology_str)
def get_open_ports(self, node_group):
handler = self.version_factory.get_version_handler(
node_group.cluster.hadoop_version)
return handler.get_open_ports(node_group)
class AmbariInfo(object):
def __init__(self, host, port, user, password):
self.host = host
self.port = port
self.user = user
self.password = password
def get_address(self):
return '{0}:{1}'.format(self.host.management_ip, self.port)
def is_ambari_info(self):
pass
def get_cluster(self):
sahara_instance = self.host.sahara_instance
return sahara_instance.cluster
def get_event_info(self):
return self.host.sahara_instance