def start_cluster(self, cluster):
nn_instance = utils.get_namenode(cluster)
datanodes = utils.get_datanodes(cluster)
jt_instance = utils.get_jobtracker(cluster)
tasktrackers = utils.get_tasktrackers(cluster)
oozie = utils.get_oozie(cluster)
with remote.get_remote(nn_instance) as r:
run.format_namenode(r)
run.start_process(r, "namenode")
snns = utils.get_secondarynamenodes(cluster)
if snns:
for snn in snns:
run.start_process(remote.get_remote(snn), "secondarynamenode")
for dn in datanodes:
run.start_process(remote.get_remote(dn), "datanode")
LOG.info("HDFS service at '%s' has been started", nn_instance.hostname)
if jt_instance:
run.start_process(remote.get_remote(jt_instance), "jobtracker")
for tt in tasktrackers:
run.start_process(remote.get_remote(tt), "tasktracker")
LOG.info("MapReduce service at '%s' has been started",
jt_instance.hostname)
if oozie:
with remote.get_remote(oozie) as r:
run.oozie_share_lib(r, nn_instance.hostname)
run.start_oozie(r)
LOG.info("Oozie service at '%s' has been started",
nn_instance.hostname)
LOG.info('Cluster %s has been started successfully' % cluster.name)
self._set_cluster_info(cluster)
def start_cluster(self, cluster):
nn_instance = utils.get_namenode(cluster)
datanodes = utils.get_datanodes(cluster)
jt_instance = utils.get_jobtracker(cluster)
tasktrackers = utils.get_tasktrackers(cluster)
oozie = utils.get_oozie(cluster)
hive_server = utils.get_hiveserver(cluster)
with remote.get_remote(nn_instance) as r:
run.format_namenode(r)
run.start_process(r, "namenode")
snns = utils.get_secondarynamenodes(cluster)
if snns:
for snn in snns:
run.start_process(remote.get_remote(snn), "secondarynamenode")
for dn in datanodes:
run.start_process(remote.get_remote(dn), "datanode")
LOG.info("HDFS service at '%s' has been started",
nn_instance.hostname)
if jt_instance:
run.start_process(remote.get_remote(jt_instance), "jobtracker")
for tt in tasktrackers:
run.start_process(remote.get_remote(tt), "tasktracker")
LOG.info("MapReduce service at '%s' has been started",
jt_instance.hostname)
if oozie:
with remote.get_remote(oozie) as r:
if c_helper.is_mysql_enable(cluster):
run.mysql_start(r, oozie)
run.oozie_create_db(r)
run.oozie_share_lib(r, nn_instance.hostname)
run.start_oozie(r)
LOG.info("Oozie service at '%s' has been started",
nn_instance.hostname)
if hive_server:
with remote.get_remote(nn_instance) as r:
run.hive_create_warehouse_dir(r)
if c_helper.is_mysql_enable(cluster):
with remote.get_remote(hive_server) as h:
if not oozie or hive_server.hostname != oozie.hostname:
run.mysql_start(h, hive_server)
run.hive_create_db(h)
run.hive_metastore_start(h)
LOG.info("Hive Metastore server at %s has been started",
hive_server.hostname)
LOG.info('Cluster %s has been started successfully' % cluster.name)
self._set_cluster_info(cluster)
def start_cluster(self, cluster):
instances = utils.get_instances(cluster)
nn_instance = utils.get_namenode(cluster)
jt_instance = utils.get_jobtracker(cluster)
oozie = utils.get_oozie(cluster)
hive_server = utils.get_hiveserver(cluster)
with remote.get_remote(nn_instance) as r:
run.format_namenode(r)
run.start_processes(r, "namenode")
for snn in utils.get_secondarynamenodes(cluster):
run.start_processes(remote.get_remote(snn), "secondarynamenode")
if jt_instance:
run.start_processes(remote.get_remote(jt_instance), "jobtracker")
self._start_tt_dn_processes(instances)
LOG.info("Hadoop services in cluster %s have been started" %
cluster.name)
if oozie:
with remote.get_remote(oozie) as r:
if c_helper.is_mysql_enable(cluster):
run.mysql_start(r, oozie)
run.oozie_create_db(r)
run.oozie_share_lib(r, nn_instance.hostname)
run.start_oozie(r)
LOG.info("Oozie service at '%s' has been started",
nn_instance.hostname)
if hive_server:
with remote.get_remote(nn_instance) as r:
run.hive_create_warehouse_dir(r)
if c_helper.is_mysql_enable(cluster):
with remote.get_remote(hive_server) as h:
if not oozie or hive_server.hostname != oozie.hostname:
run.mysql_start(h, hive_server)
run.hive_create_db(h)
run.hive_metastore_start(h)
LOG.info("Hive Metastore server at %s has been started",
hive_server.hostname)
LOG.info('Cluster %s has been started successfully' % cluster.name)
self._set_cluster_info(cluster)
def start_cluster(self, cluster):
nn_instance = utils.get_namenode(cluster)
with remote.get_remote(nn_instance) as r:
run.format_namenode(r)
run.start_processes(r, "namenode")
for snn in utils.get_secondarynamenodes(cluster):
run.start_processes(remote.get_remote(snn), "secondarynamenode")
jt_instance = utils.get_jobtracker(cluster)
if jt_instance:
run.start_processes(remote.get_remote(jt_instance), "jobtracker")
self._start_tt_dn_processes(utils.get_instances(cluster))
self._await_datanodes(cluster)
LOG.info("Hadoop services in cluster %s have been started" %
cluster.name)
oozie = utils.get_oozie(cluster)
if oozie:
with remote.get_remote(oozie) as r:
if c_helper.is_mysql_enable(cluster):
run.mysql_start(r, oozie)
run.oozie_create_db(r)
run.oozie_share_lib(r, nn_instance.hostname())
run.start_oozie(r)
LOG.info("Oozie service at '%s' has been started",
nn_instance.hostname())
hive_server = utils.get_hiveserver(cluster)
if hive_server:
with remote.get_remote(nn_instance) as r:
run.hive_create_warehouse_dir(r)
if c_helper.is_mysql_enable(cluster):
with remote.get_remote(hive_server) as h:
if not oozie or hive_server.hostname() != oozie.hostname():
run.mysql_start(h, hive_server)
run.hive_create_db(h)
run.hive_metastore_start(h)
LOG.info("Hive Metastore server at %s has been started",
hive_server.hostname())
LOG.info('Cluster %s has been started successfully' % cluster.name)
self._set_cluster_info(cluster)
def _configure_services(client, cluster):
nn_host = u.get_namenode(cluster).fqdn()
snn = u.get_secondarynamenodes(cluster)
snn_host = snn[0].fqdn() if snn else None
jt_host = u.get_jobtracker(cluster).fqdn() if u.get_jobtracker(
cluster) else None
dn_hosts = [dn.fqdn() for dn in u.get_datanodes(cluster)]
tt_hosts = [tt.fqdn() for tt in u.get_tasktrackers(cluster)]
oozie_host = u.get_oozie(cluster).fqdn() if u.get_oozie(cluster) else None
hive_host = u.get_hiveserver(cluster).fqdn() if u.get_hiveserver(
cluster) else None
services = []
if u.get_namenode(cluster):
services += ['hdfs']
if u.get_jobtracker(cluster):
services += ['mapred']
if oozie_host:
services += ['oozie']
services += ['pig']
if hive_host:
services += ['hive']
LOG.debug("Add services: %s" % ', '.join(services))
client.services.add(services)
LOG.debug("Assign roles to hosts")
client.services.hdfs.add_nodes('PrimaryNameNode', [nn_host])
client.services.hdfs.add_nodes('DataNode', dn_hosts)
if snn:
client.services.hdfs.add_nodes('SecondaryNameNode', [snn_host])
if oozie_host:
client.services.oozie.add_nodes('Oozie', [oozie_host])
if hive_host:
client.services.hive.add_nodes('HiveServer', [hive_host])
if jt_host:
client.services.mapred.add_nodes('JobTracker', [jt_host])
client.services.mapred.add_nodes('TaskTracker', tt_hosts)
def _configure_services(client, cluster):
nn_host = u.get_namenode(cluster).fqdn()
snn = u.get_secondarynamenodes(cluster)
snn_host = snn[0].fqdn() if snn else None
jt_host = u.get_jobtracker(cluster).fqdn()
dn_hosts = [dn.fqdn() for dn in u.get_datanodes(cluster)]
tt_hosts = [tt.fqdn() for tt in u.get_tasktrackers(cluster)]
oozie_host = u.get_oozie(cluster).fqdn() if u.get_oozie(
cluster) else None
hive_host = u.get_hiveserver(cluster).fqdn() if u.get_hiveserver(
cluster) else None
services = []
if u.get_namenode(cluster):
services += ['hdfs']
if u.get_jobtracker(cluster):
services += ['mapred']
if oozie_host:
services += ['oozie']
services += ['pig']
if hive_host:
services += ['hive']
LOG.debug("Add services: %s" % ', '.join(services))
client.services.add(services)
LOG.debug("Assign roles to hosts")
client.services.hdfs.add_nodes('PrimaryNameNode', [nn_host])
client.services.hdfs.add_nodes('DataNode', dn_hosts)
if snn:
client.services.hdfs.add_nodes('SecondaryNameNode', [snn_host])
if oozie_host:
client.services.oozie.add_nodes('Oozie', [oozie_host])
if hive_host:
client.services.hive.add_nodes('HiveServer', [hive_host])
client.services.mapred.add_nodes('JobTracker', [jt_host])
client.services.mapred.add_nodes('TaskTracker', tt_hosts)
def start_cluster(self, cluster):
nn_instance = utils.get_namenode(cluster)
datanodes = utils.get_datanodes(cluster)
jt_instance = utils.get_jobtracker(cluster)
tasktrackers = utils.get_tasktrackers(cluster)
oozie = utils.get_oozie(cluster)
with remote.get_remote(nn_instance) as r:
run.format_namenode(r)
run.start_process(r, "namenode")
snns = utils.get_secondarynamenodes(cluster)
if snns:
for snn in snns:
run.start_process(remote.get_remote(snn), "secondarynamenode")
for dn in datanodes:
run.start_process(remote.get_remote(dn), "datanode")
LOG.info("HDFS service at '%s' has been started",
nn_instance.hostname)
if jt_instance:
run.start_process(remote.get_remote(jt_instance), "jobtracker")
for tt in tasktrackers:
run.start_process(remote.get_remote(tt), "tasktracker")
LOG.info("MapReduce service at '%s' has been started",
jt_instance.hostname)
if oozie:
with remote.get_remote(oozie) as r:
run.oozie_share_lib(r, nn_instance.hostname)
run.start_oozie(r)
LOG.info("Oozie service at '%s' has been started",
nn_instance.hostname)
LOG.info('Cluster %s has been started successfully' % cluster.name)
self._set_cluster_info(cluster)
