def updateJobProperties(cls,
propFile,
properties=None,
haEnabled=False,
debug=False):
fileSystemName = Hadoop.getFSDefaultValue()
jobTrackerIP = MAPRED.getJobtrackerAddress()
jobTracker = jobTrackerIP[0] + ":" + jobTrackerIP[1]
if not properties:
properties = {}
if not properties.has_key('nameNode'):
properties['nameNode'] = fileSystemName
if not properties.has_key('jobTracker'):
properties['jobTracker'] = jobTracker
if "hcatalog" in propFile:
if Hadoop.isSecure():
kerberosPrincipal = Hive.getConfigValue(
"hive.metastore.kerberos.principal")
properties[
'hive.metastore.kerberos.principal'] = kerberosPrincipal
logger.info("Updating for hcatalog workflow")
hcatNode = Hive.getConfigValue("hive.metastore.uris").replace(
'thrift', 'hcat')
logger.info("Hcat node is " + hcatNode)
properties['hcatNode'] = hcatNode
if Hadoop.isSecure():
# determine the namenode and the jobtracker principal
nnPrincipal = None
if haEnabled:
nnPrincipal = HDFS.getNameNodePrincipal().replace(
'_HOST', HDFS.getNamenodeByState('active'))
else:
nnPrincipal = HDFS.getNameNodePrincipal().replace(
'_HOST',
HDFS.getNamenodeHttpAddress()[0])
jtPrincipal = MAPRED.getMasterPrincipal().replace(
'_HOST', jobTrackerIP[0])
properties['dfs.namenode.kerberos.principal'] = nnPrincipal
properties['mapreduce.jobtracker.kerberos.principal'] = jtPrincipal
wfPath = util.getPropertyValueFromFile(propFile,
"oozie.wf.application.path")
if wfPath != None and wfPath.find("hdfs://localhost:9000") != -1:
wfPath = wfPath.replace("hdfs://localhost:9000", fileSystemName)
logger.info("Value of replaced oozie.wf.application.path is " +
wfPath)
properties['oozie.wf.application.path'] = wfPath
util.writePropertiesToFile(propFile, propFile, properties)
if debug:
logger.info('Content of properties file %s' % propFile)
f = open(propFile, 'r')
# print the file to the console
logger.info(f.read())
f.close()
def getDatabaseFlavor(cls):
dbdriver = Hive.getConfigValue("javax.jdo.option.ConnectionDriverName")
if ("oracle" in dbdriver):
return "oracle"
elif ("postgresql" in dbdriver):
dbUrl = Hive.getConfigValue("javax.jdo.option.ConnectionURL")
m = re.search('jdbc:postgresql://(.*):.*', dbUrl)
dbHost = Machine.getfqdn()
if m and m.group(1):
dbHost = m.group(1)
dbVersion = Machine.getDBVersion('postgres', host=dbHost)
if dbVersion:
return "postgres-%s" % dbVersion
else:
return "postgres"
elif ("derby" in dbdriver):
return "derby"
elif ("mysql" in dbdriver):
return "mysql"
return ""
def perform_post_upgrade_steps(self):
if Config.getEnv("HDP_STACK_INSTALLED").lower() == "true":
from beaver.component.hadoop import Hadoop, HDFS
from beaver.component.hive import Hive
COMPONENT = str(self.COMPONENT)
HDFS_USER = Config.get('hadoop', 'HDFS_USER')
if 'experiment' in COMPONENT and Hive.isInstalled():
HIVE_WAREHOUSE_DIR = Hive.getConfigValue(
"hive.metastore.warehouse.dir", defaultValue="/apps/hive/warehouse"
)
HDFS.chmod(HDFS_USER, 777, HIVE_WAREHOUSE_DIR, True)
else:
UpgradeLogger.reportProgress("No additional post-upgrade steps defined for EU", True)
else:
logger.info("No additional post-upgrade steps defined for EU on HDF")
def setup_storm_hive_topology(cls, useStandaloneCmd):
from beaver.component.hive import Hive
storm_version = Storm.getVersion(useStandaloneCmd=True)
hive_version = Hive.getVersion()
HIVE_METASTORE_URI = Hive.getConfigValue(
"hive.metastore.uris", defaultValue="thrift://localhost:9083")
global HIVE_METASTORE_URI
global HIVE_HOST
global HIVE_PORT
global HIVE_WAREHOUSE_DIR
HIVE_WAREHOUSE_DIR = Hive.getConfigValue(
"hive.metastore.warehouse.dir",
defaultValue="/apps/hive/warehouse")
HIVE_HOST = Hive.getHiveHost()
HIVE_PORT = Hive.getMetastoreThriftPort()
if Storm.isDalorBeyond():
JAVA_HIVE_SRC_DIR = os.path.join(Config.getEnv('WORKSPACE'),
'tests', 'rolling_upgrade',
'Storm', '2_3', 'storm-hive',
'java')
else:
JAVA_HIVE_SRC_DIR = os.path.join(Config.getEnv('WORKSPACE'),
'tests', 'rolling_upgrade',
'Storm', '2_2', 'storm-hive',
'java')
# hive.txn.manager and hive.support.concurrency are set through ambari as per bug-40500
#logger.info("Restart Hive")
#changes = {'hive-site.xml': {'hive.txn.manager': 'org.apache.hadoop.hive.ql.lockmgr.DbTxnManager',
# 'hive.support.concurrency': 'true'}}
#Hive.modifyConfig(changes, services=['metastore'], restartService=True)
logger.info("Create test database in Hive")
exit_code, stdout = Hive.runQuery(
cls.get_set_queue_cmd(useStandaloneCmd) +
" drop database if exists stormdb cascade; \
create database stormdb;")
ruAssert("Storm", exit_code == 0,
"[StormHiveSetup] Failed to create test database" + stdout)
HDFS.chmod(runasUser=HDFS.getHDFSUser(),
perm=777,
directory=HIVE_WAREHOUSE_DIR + "/" + DATABASE_NAME + ".db")
#copy tests/storm/storm-hive/java to artifacts/storm-hive-tests
logger.info("JAVA_SRC_DIR " + JAVA_HIVE_SRC_DIR)
logger.info("LOCAL_WORK_DIR " + LOCAL_HIVE_WORK_DIR)
Machine.copy(JAVA_HIVE_SRC_DIR,
LOCAL_HIVE_WORK_DIR,
user=None,
passwd=None)
#mvn package
if Machine.isWindows():
(_, _) = Maven.run(
'package -D%s=%s -D%s=%s -D%s=%s -D%s=%s' %
(HADOOP_VERSION_MAVEN_PARAMETER, HADOOP_VERSION,
STORM_VERSION_MAVEN_PARAMETER, storm_version,
HIVE_VERSION_MAVEN_PARAMETER,
hive_version, PUBLIC_REPO_MAVEN_PARAMETER,
Maven.getPublicRepoUrl(), CORE_FILE_MAVEN_PARAMETER,
CORE_FILE, HADOOP_CORE_MAVEN_PARAMETER, HADOOP_CONF,
HIVE_CORE_MAVEN_PARAMETER, HIVE_CORE_DIR,
HIVE_FILE_MAVEN_PARAMETER, HIVE_FILE),
cwd=LOCAL_HIVE_WORK_DIR)
else:
(_, _) = Maven.run('package',
cwd=LOCAL_HIVE_WORK_DIR,
env={
HADOOP_VERSION_MAVEN_PARAMETER:
HADOOP_VERSION,
STORM_VERSION_MAVEN_PARAMETER:
storm_version,
HIVE_VERSION_MAVEN_PARAMETER:
hive_version,
PUBLIC_REPO_MAVEN_PARAMETER:
Maven.getPublicRepoUrl(),
CORE_FILE_MAVEN_PARAMETER:
CORE_FILE,
HADOOP_CONF_MAVEN_PARAMETER:
HADOOP_CONF,
HDFS_FILE_MAVEN_PARAMETER:
HDFS_FILE,
HADOOP_CORE_MAVEN_PARAMETER:
HADOOP_CONF,
HIVE_CORE_MAVEN_PARAMETER:
HIVE_CORE_DIR,
HIVE_FILE_MAVEN_PARAMETER:
HIVE_FILE
})
create_table_q = "use %s; \
drop table if exists %s; \
create table %s (id int, name string, phone string, street string) \
partitioned by (city string, state string) \
clustered by (id) into %s buckets \
stored as orc \
tblproperties ('transactional'='true');" % (
DATABASE_NAME, HIVE_TABLE_NAME, HIVE_TABLE_NAME, "5")
exit_code, stdout = Hive.runQuery(
cls.get_set_queue_cmd(useStandaloneCmd) + create_table_q)
ruAssert(
"Storm", exit_code == 0,
"[StormHiveSetup] Failed to create test table userdata_partitioned"
)
HDFS.chmod(runasUser=HDFS.getHDFSUser(),
perm=777,
directory=HIVE_WAREHOUSE_DIR + "/" + DATABASE_NAME +
".db/" + HIVE_TABLE_NAME)
