def runCassandra(status='stop'):
with settings(warn_only=True):
fabfile.__normalUser()
cassandra_path = os.path.join(
'/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
print cassandra_path
if (status == 'start'):
if (not (env.host in fabfile.cf.get(fabfile.activeSession,
'admin_ip'))
and not (env.host in fabfile.cf.get(
'cassandra', 'one_seed_ip'))):
out = run(
'nohup ' +
os.path.join(cassandra_path,
'bin/cassandra -p cassandraPID && sleep 5') +
' ',
pty=True,
combine_stderr=True)
print out
elif (status == 'stop'):
if (not (env.host in fabfile.cf.get(fabfile.activeSession,
'admin_ip'))):
pid = run('cat ' +
os.path.join('/home', env.user, 'cassandraPID'))
run("kill -9 " + pid)
run('rm ' + os.path.join('/home', env.user, 'cassandraPID'))
else:
print 'unknow command ' + status + ", only support start or stop"
def runOneCassandraSeed():
if (env.host in fabfile.cf.get('cassandra','one_seed_ip')):
fabfile.__normalUser()
cassandra_path=os.path.join('/home',env.user,fabfile.cf.get('cassandra','cassandra_folder'))
print cassandra_path
out=run('nohup ' + os.path.join(cassandra_path,'bin/cassandra -p cassandraPID')+' ',pty=True, combine_stderr=True)
print out
def modifyCassandra():
if (not (env.host in fabfile.cf.get(fabfile.activeSession, 'admin_ip'))):
fabfile.__normalUser()
cassandra_path = os.path.join(
'/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
yaml = os.path.join(
os.path.split(env.real_fabfile)[0], 'files/cassandra.yaml')
put(yaml, os.path.join(cassandra_path, 'conf/cassandra.yaml'))
def formatHadoop():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf=fabfile.cf
special = cf.get('hadoop', 'master_public_ip')
hadoop_folder=os.path.join('/home',env.user,cf.get('hadoop','hadoop_folder'))
with cd(hadoop_folder):
if(env.host==special):
run('bin/hdfs namenode -format '+ cf.get("hadoop",'format_cluster_name'))
def showCassandra():
if (env.host in fabfile.cf.get('cassandra', 'one_seed_ip')):
fabfile.__normalUser()
cassandra_path = os.path.join(
'/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
out = run(os.path.join(cassandra_path, 'bin/nodetool status'),
pty=True,
combine_stderr=True)
print out
def stopSpark2_4_0():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf=fabfile.cf
special = cf.get('spark', 'master_public_ip')
spark_path = os.path.join('/home', env.user, cf.get('spark', 'spark_folder'))
with cd(spark_path):
if(env.host==special):
run('sbin/stop-all.sh')
def stopSpark2_4_0():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf = fabfile.cf
special = cf.get('spark', 'master_public_ip')
spark_path = os.path.join('/home', env.user,
cf.get('spark', 'spark_folder'))
with cd(spark_path):
if (env.host == special):
run('sbin/stop-all.sh')
def stopHadoop():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf=fabfile.cf
special = cf.get('hadoop', 'master_public_ip')
hadoop_folder=os.path.join('/home',env.user,cf.get('hadoop','hadoop_folder'))
with cd(hadoop_folder):
if(env.host==special):
run('sbin/stop-all.sh')
run('sbin/mr-jobhistory-daemon.sh stop historyserver')
def stopHadoop():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf = fabfile.cf
special = cf.get('hadoop', 'master_public_ip')
hadoop_folder = os.path.join('/home', env.user,
cf.get('hadoop', 'hadoop_folder'))
with cd(hadoop_folder):
if (env.host == special):
run('sbin/stop-all.sh')
run('sbin/mr-jobhistory-daemon.sh stop historyserver')
def formatHadoop():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf = fabfile.cf
special = cf.get('hadoop', 'master_public_ip')
hadoop_folder = os.path.join('/home', env.user,
cf.get('hadoop', 'hadoop_folder'))
with cd(hadoop_folder):
if (env.host == special):
run('bin/hdfs namenode -format ' +
cf.get("hadoop", 'format_cluster_name'))
def startSpark2_4_0():
with settings(prompts={
'Are you sure you want to continue connecting (yes/no)?':'yes'
}):
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf=fabfile.cf
special = cf.get('spark', 'master_public_ip')
spark_path = os.path.join('/home', env.user, cf.get('spark', 'spark_folder'))
with cd(spark_path):
if(env.host==special):
run('sbin/start-all.sh')
def runOneCassandraSeed():
if (env.host in fabfile.cf.get('cassandra', 'one_seed_ip')):
fabfile.__normalUser()
cassandra_path = os.path.join(
'/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
print cassandra_path
out = run(
'nohup ' +
os.path.join(cassandra_path, 'bin/cassandra -p cassandraPID') +
' ',
pty=True,
combine_stderr=True)
print out
def startHadoop():
with settings(prompts={
'Are you sure you want to continue connecting (yes/no)?':'yes'
}):
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf=fabfile.cf
special = cf.get('hadoop', 'master_public_ip')
hadoop_folder=os.path.join('/home',env.user,cf.get('hadoop','hadoop_folder'))
with cd(hadoop_folder):
if(env.host==special):
run('sbin/start-all.sh')
run('sbin/mr-jobhistory-daemon.sh start historyserver')
def distributeSpark2_4_0():
if (not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts:
fabfile.__normalUser()
print 'will create a temp file /home/username/fabric-spark.tar'
cf = fabfile.cf
spark_path = os.path.join('/home', env.user, cf.get('spark', 'spark_folder'))
spark_config_folder = os.path.join('/home', env.user, cf.get('spark', 'spark_folder'), 'conf')
put(os.path.join(os.path.split(env.real_fabfile)[0], cf.get('spark', 'spark_file')), os.path.join('/home', env.user, 'fabric-spark.tar'))
run('tar -xf ' + os.path.join('/home', env.user, 'fabric-spark.tar'))
if not remoteFileExist(cf.get('spark', 'spark_work')):
run('mkdir -p ' + cf.get('spark', 'spark_work'))
with cd(spark_config_folder):
print '清空slaves...'
put(os.path.join(os.path.split(env.real_fabfile)[0], 'files/spark/slaves'), spark_config_folder)
# run("cp slaves.template > slaves")
run("cat /dev/null > slaves")
print '填写slaves...'
# special = cf.get('spark', 'master_ip')
for node in cf.get('spark', 'slaves').split(","):
run("echo " + node + ">> slaves")
print '填写spark-env.sh'
put(os.path.join(os.path.split(env.real_fabfile)[0], 'files/spark/spark-env.sh'), spark_config_folder)
# SPARK_MASTER_PORT
run("echo 'SPARK_MASTER_PORT=7077' >> spark-env.sh")
# SPARK_MASTER_HOST
run("echo 'SPARK_MASTER_HOST=" + cf.get('spark', 'master_ip') + "' >> spark-env.sh")
# SPARK_LOCAL_IP
localippara1 = 'ifconfig -a|grep inet|grep -v inet6|grep -v 127|awk '
localippara2 = '\'{print $2}\''
localippara3 = '|tr -d '
localippara4 = '\"addr:\"'
localippara5 = '|sed -n '
localippara6 = '\'1p\''
# SparkLocalIp = run('ifconfig -a|grep inet|grep -v inet6|grep -v 127|awk '{print $2}'|tr -d "addr:"|sed -n '1p'')
SparkLocalIp = localippara1 + localippara2 + localippara3 + localippara4 + localippara5 + localippara6
out = run(SparkLocalIp)
run("echo 'SPARK_LOCAL_IP=" + out + "' >> spark-env.sh")
# SPARK_HOME
run("echo 'SPARK_HOME=" + spark_path + "' >> spark-env.sh")
# JAVA_HOME
jdk_path=os.path.join('/home/', env.user, cf.get(fabfile.activeSession, 'jdk_folder'))
run("echo 'JAVA_HOME=" + jdk_path + "' >> spark-env.sh")
# SPARK_WORK_DIR
run("echo 'SPARK_WORK_DIR=" + cf.get('spark', 'spark_work') +"' >> spark-env.sh")
# SPARK_WORKER_OPTS
spark_work_opts='"-Dspark.worker.cleanup.enabled=true -Dspark.worker.cleanup.interval=1800 -Dspark.worker.cleanup.appDataTtl=3600"'
run("echo 'SPARK_WORKER_OPTS=" + spark_work_opts + "' >> spark-env.sh")
def distributeCassandra():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
print 'will create a temp file /home/username/fabric-cassandra.tar.gz'
cf=fabfile.cf
put(os.path.join(os.path.split(env.real_fabfile)[0], cf.get('cassandra','cassandra_file')), os.path.join('/home',env.user,'fabric-cassandra.tar.gz'))
run('tar -xzf '+ os.path.join('/home',env.user,'fabric-cassandra.tar.gz'))
cassandra_path=os.path.join('/home',env.user,cf.get('cassandra','cassandra_folder'))
yaml=os.path.join(os.path.split(env.real_fabfile)[0], 'files/cassandra.yaml')
logback=os.path.join(os.path.split(env.real_fabfile)[0], 'files/cassandra-logback.xml')
put(yaml,os.path.join(cassandra_path,'conf/cassandra.yaml'))
put(logback, os.path.join(cassandra_path, 'conf/logback.yaml'))
run('mkdir '+ cf.get('cassandra','data_folder'))
def startSpark2_4_0():
with settings(
prompts={
'Are you sure you want to continue connecting (yes/no)?': 'yes'
}):
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf = fabfile.cf
special = cf.get('spark', 'master_public_ip')
spark_path = os.path.join('/home', env.user,
cf.get('spark', 'spark_folder'))
with cd(spark_path):
if (env.host == special):
run('sbin/start-all.sh')
def startHadoop():
with settings(
prompts={
'Are you sure you want to continue connecting (yes/no)?': 'yes'
}):
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
cf = fabfile.cf
special = cf.get('hadoop', 'master_public_ip')
hadoop_folder = os.path.join('/home', env.user,
cf.get('hadoop', 'hadoop_folder'))
with cd(hadoop_folder):
if (env.host == special):
run('sbin/start-all.sh')
run('sbin/mr-jobhistory-daemon.sh start historyserver')
def runCassandra(status='stop'):
with settings(warn_only=True):
fabfile.__normalUser()
cassandra_path = os.path.join('/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
print cassandra_path
if (status == 'start'):
if (not (env.host in fabfile.cf.get(fabfile.activeSession,'admin_ip')) and not (env.host in fabfile.cf.get('cassandra','one_seed_ip'))):
out=run('nohup ' + os.path.join(cassandra_path,'bin/cassandra -p cassandraPID && sleep 5')+' ',pty=True, combine_stderr=True)
print out
elif(status=='stop'):
if (not (env.host in fabfile.cf.get(fabfile.activeSession, 'admin_ip'))):
pid=run('cat '+os.path.join('/home',env.user,'cassandraPID'))
run("kill -9 "+pid)
run('rm '+os.path.join('/home',env.user,'cassandraPID'))
else:
print 'unknow command '+ status+", only support start or stop"
def distributeHadoop2_8_5():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
print 'will create a temp file /home/username/fabric-hadoop.tar.gz'
cf=fabfile.cf
put(os.path.join(os.path.split(env.real_fabfile)[0], cf.get('hadoop','hadoop_file')), os.path.join('/home',env.user,'fabric-hadoop.tar.gz'))
run('tar -xzf '+ os.path.join('/home',env.user,'fabric-hadoop.tar.gz'))
if not remoteFileExist(cf.get('hadoop','data_folder')):
run('mkdir -p ' + cf.get('hadoop','data_folder'))
hadoop_config_folder=os.path.join('/home',env.user,cf.get('hadoop','hadoop_folder'),'etc/hadoop')
with cd(hadoop_config_folder):
modifyJDK="sed -i 's/export JAVA_HOME=.*/export JAVA_HOME=" + os.path.join('/home/',env.user, cf.get(fabfile.activeSession,'jdk_folder')).replace("/","\\/") + "/g' hadoop-env.sh"
run(modifyJDK)
print '清空slaves...'
run("cat /dev/null > slaves")
print '填写slaves...'
special = cf.get('hadoop','master_ip')
for node in cf.get('hadoop','slaves').split(","):
#if(not (node == special and cf.get('hadoop','master_as_a_slave')=="0")):
run("echo "+ node + ">> slaves")
print '填写mapred-site.xml'
put(os.path.join(os.path.split(env.real_fabfile)[0], 'files/hadoop/mapred-site.xml'),hadoop_config_folder)
modify="sed -i 's/<value>MASTERIP:10020.*/"+"<value>"+special+":10020<\\/value>"+"/g' mapred-site.xml"
run(modify)
modify = "sed -i 's/<value>MASTERIP:19888.*/" + "<value>" + special + ":19888<\\/value>" + "/g' mapred-site.xml"
run(modify)
print '填写core-site.xml'
put(os.path.join(os.path.split(env.real_fabfile)[0], 'files/hadoop/core-site.xml'), hadoop_config_folder)
modifyFSURL="sed -i 's/<value>.*/"+"<value>hdfs:\\/\\/"+special+":9000<\\/value>"+"/g' core-site.xml"
run(modifyFSURL)
print '填写yarn-site.xml'
put(os.path.join(os.path.split(env.real_fabfile)[0], 'files/hadoop/yarn-site.xml'), hadoop_config_folder)
modify = "sed -i 's/<value>MASTERIP:8031.*/" + "<value>" + special + ":8031<\\/value>" + "/g' yarn-site.xml"
run(modify)
modify = "sed -i 's/<value>MASTERIP:8032.*/" + "<value>" + special + ":8032<\\/value>" + "/g' yarn-site.xml"
run(modify)
modify = "sed -i 's/<value>MASTERIP:8030.*/" + "<value>" + special + ":8030<\\/value>" + "/g' yarn-site.xml"
run(modify)
print '填写hdfs-site.xml'
put(os.path.join(os.path.split(env.real_fabfile)[0], 'files/hadoop/hdfs-site.xml'), hadoop_config_folder)
modify="sed -i 's/<value>DATADIR.*/<value>" + cf.get('hadoop','data_folder').replace("/", "\\/") +"<\\/value>/g' hdfs-site.xml"
run(modify)
def distributeCassandra():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
print 'will create a temp file /home/username/fabric-cassandra.tar.gz'
cf = fabfile.cf
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
cf.get('cassandra', 'cassandra_file')),
os.path.join('/home', env.user, 'fabric-cassandra.tar.gz'))
run('tar -xzf ' +
os.path.join('/home', env.user, 'fabric-cassandra.tar.gz'))
cassandra_path = os.path.join('/home', env.user,
cf.get('cassandra', 'cassandra_folder'))
yaml = os.path.join(
os.path.split(env.real_fabfile)[0], 'files/cassandra.yaml')
logback = os.path.join(
os.path.split(env.real_fabfile)[0], 'files/cassandra-logback.xml')
put(yaml, os.path.join(cassandra_path, 'conf/cassandra.yaml'))
put(logback, os.path.join(cassandra_path, 'conf/logback.yaml'))
run('mkdir ' + cf.get('cassandra', 'data_folder'))
def showCassandra():
if (env.host in fabfile.cf.get('cassandra', 'one_seed_ip')):
fabfile.__normalUser()
cassandra_path = os.path.join('/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
out = run(os.path.join(cassandra_path, 'bin/nodetool status'), pty=True, combine_stderr=True)
print out
def rmCassandraData(status='stop'):
if (not (env.host in fabfile.cf.get(fabfile.activeSession,'admin_ip'))):
fabfile.__normalUser()
run('rm -rf '+fabfile.cf.get('cassandra','data_folder'))
run('mkdir ' + fabfile.cf.get('cassandra', 'data_folder'))
def distributeSpark2_4_0():
if (not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts:
fabfile.__normalUser()
print 'will create a temp file /home/username/fabric-spark.tar'
cf = fabfile.cf
spark_path = os.path.join('/home', env.user,
cf.get('spark', 'spark_folder'))
spark_config_folder = os.path.join('/home', env.user,
cf.get('spark', 'spark_folder'),
'conf')
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
cf.get('spark', 'spark_file')),
os.path.join('/home', env.user, 'fabric-spark.tar'))
run('tar -xf ' + os.path.join('/home', env.user, 'fabric-spark.tar'))
if not remoteFileExist(cf.get('spark', 'spark_work')):
run('mkdir -p ' + cf.get('spark', 'spark_work'))
with cd(spark_config_folder):
print '清空slaves...'
put(
os.path.join(
os.path.split(env.real_fabfile)[0], 'files/spark/slaves'),
spark_config_folder)
# run("cp slaves.template > slaves")
run("cat /dev/null > slaves")
print '填写slaves...'
# special = cf.get('spark', 'master_ip')
for node in cf.get('spark', 'slaves').split(","):
run("echo " + node + ">> slaves")
print '填写spark-env.sh'
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
'files/spark/spark-env.sh'), spark_config_folder)
# SPARK_MASTER_PORT
run("echo 'SPARK_MASTER_PORT=7077' >> spark-env.sh")
# SPARK_MASTER_HOST
run("echo 'SPARK_MASTER_HOST=" + cf.get('spark', 'master_ip') +
"' >> spark-env.sh")
# SPARK_LOCAL_IP
localippara1 = 'ifconfig -a|grep inet|grep -v inet6|grep -v 127|awk '
localippara2 = '\'{print $2}\''
localippara3 = '|tr -d '
localippara4 = '\"addr:\"'
localippara5 = '|sed -n '
localippara6 = '\'1p\''
# SparkLocalIp = run('ifconfig -a|grep inet|grep -v inet6|grep -v 127|awk '{print $2}'|tr -d "addr:"|sed -n '1p'')
SparkLocalIp = localippara1 + localippara2 + localippara3 + localippara4 + localippara5 + localippara6
out = run(SparkLocalIp)
run("echo 'SPARK_LOCAL_IP=" + out + "' >> spark-env.sh")
# SPARK_HOME
run("echo 'SPARK_HOME=" + spark_path + "' >> spark-env.sh")
# JAVA_HOME
jdk_path = os.path.join(
'/home/', env.user, cf.get(fabfile.activeSession,
'jdk_folder'))
run("echo 'JAVA_HOME=" + jdk_path + "' >> spark-env.sh")
# SPARK_WORK_DIR
run("echo 'SPARK_WORK_DIR=" + cf.get('spark', 'spark_work') +
"' >> spark-env.sh")
# SPARK_WORKER_OPTS
spark_work_opts = '"-Dspark.worker.cleanup.enabled=true -Dspark.worker.cleanup.interval=1800 -Dspark.worker.cleanup.appDataTtl=3600"'
run("echo 'SPARK_WORKER_OPTS=" + spark_work_opts +
"' >> spark-env.sh")
def modifyCassandra():
if (not (env.host in fabfile.cf.get(fabfile.activeSession,'admin_ip'))):
fabfile.__normalUser()
cassandra_path = os.path.join('/home', env.user, fabfile.cf.get('cassandra', 'cassandra_folder'))
yaml = os.path.join(os.path.split(env.real_fabfile)[0], 'files/cassandra.yaml')
put(yaml, os.path.join(cassandra_path, 'conf/cassandra.yaml'))
def rmCassandraData(status='stop'):
if (not (env.host in fabfile.cf.get(fabfile.activeSession, 'admin_ip'))):
fabfile.__normalUser()
run('rm -rf ' + fabfile.cf.get('cassandra', 'data_folder'))
run('mkdir ' + fabfile.cf.get('cassandra', 'data_folder'))
def distributeHadoop2_8_5():
if ((not fabfile.myenv.append) or env.host in fabfile.myenv.new_hosts):
fabfile.__normalUser()
print 'will create a temp file /home/username/fabric-hadoop.tar.gz'
cf = fabfile.cf
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
cf.get('hadoop', 'hadoop_file')),
os.path.join('/home', env.user, 'fabric-hadoop.tar.gz'))
run('tar -xzf ' +
os.path.join('/home', env.user, 'fabric-hadoop.tar.gz'))
if not remoteFileExist(cf.get('hadoop', 'data_folder')):
run('mkdir -p ' + cf.get('hadoop', 'data_folder'))
hadoop_config_folder = os.path.join('/home', env.user,
cf.get('hadoop', 'hadoop_folder'),
'etc/hadoop')
with cd(hadoop_config_folder):
modifyJDK = "sed -i 's/export JAVA_HOME=.*/export JAVA_HOME=" + os.path.join(
'/home/', env.user,
cf.get(fabfile.activeSession, 'jdk_folder')).replace(
"/", "\\/") + "/g' hadoop-env.sh"
run(modifyJDK)
print '清空slaves...'
run("cat /dev/null > slaves")
print '填写slaves...'
special = cf.get('hadoop', 'master_ip')
for node in cf.get('hadoop', 'slaves').split(","):
#if(not (node == special and cf.get('hadoop','master_as_a_slave')=="0")):
run("echo " + node + ">> slaves")
print '填写mapred-site.xml'
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
'files/hadoop/mapred-site.xml'), hadoop_config_folder)
modify = "sed -i 's/<value>MASTERIP:10020.*/" + "<value>" + special + ":10020<\\/value>" + "/g' mapred-site.xml"
run(modify)
modify = "sed -i 's/<value>MASTERIP:19888.*/" + "<value>" + special + ":19888<\\/value>" + "/g' mapred-site.xml"
run(modify)
print '填写core-site.xml'
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
'files/hadoop/core-site.xml'), hadoop_config_folder)
modifyFSURL = "sed -i 's/<value>.*/" + "<value>hdfs:\\/\\/" + special + ":9000<\\/value>" + "/g' core-site.xml"
run(modifyFSURL)
print '填写yarn-site.xml'
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
'files/hadoop/yarn-site.xml'), hadoop_config_folder)
modify = "sed -i 's/<value>MASTERIP:8031.*/" + "<value>" + special + ":8031<\\/value>" + "/g' yarn-site.xml"
run(modify)
modify = "sed -i 's/<value>MASTERIP:8032.*/" + "<value>" + special + ":8032<\\/value>" + "/g' yarn-site.xml"
run(modify)
modify = "sed -i 's/<value>MASTERIP:8030.*/" + "<value>" + special + ":8030<\\/value>" + "/g' yarn-site.xml"
run(modify)
print '填写hdfs-site.xml'
put(
os.path.join(
os.path.split(env.real_fabfile)[0],
'files/hadoop/hdfs-site.xml'), hadoop_config_folder)
modify = "sed -i 's/<value>DATADIR.*/<value>" + cf.get(
'hadoop', 'data_folder').replace(
"/", "\\/") + "<\\/value>/g' hdfs-site.xml"
run(modify)
