我有一个python脚本,我想用oozie来调度它。我使用Oozie shell操作来调用脚本。脚本中有一个直线命令。当我运行oozie工作流时,出现错误“sh:beeline:command not found”。如果我从edge节点手动运行这个脚本或beeline命令,它运行得非常好。我的数据平台是HortonWorks2.6。下面是我的工作流.xml和python脚本:
工作流.xml
<workflow-app xmlns="uri:oozie:workflow:0.3" name="hive2-wf">
<credentials>
<credential name='hcat-creds' type='hcat'>
<property>
<name>hcat.metastore.uri</name>
<value>thrift://host:9083</value>
</property>
<property>
<name>hcat.metastore.principal</name>
<value>hive/_HOST@SOLON.PRD</value>
</property>
</credential>
</credentials>
<start to="python-node"/>
<action name="python-node" cred="hcat-creds">
<shell xmlns="uri:oozie:shell-action:0.1">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<exec>run_validations.py</exec>
<argument>--jdbcURL</argument><argument>${jdbcURL}</argument>
<argument>--jdbcPrincipal</argument><argument>${jdbcPrincipal}</argument>
<env-var>PYTHONPATH=/bin/python</env-var>
<env-var>PYTHON_EGG_CACHE=/tmp</env-var>
<env-var>PATH=/usr/bin</env-var>
<env-var>HADOOP_CLASSPATH=${HADOOP_CLASSPATH}</env-var>
<file>run_validations.py</file>
</shell>
<ok to="end"/>
<error to="fail"/>
</action>
<kill name="fail">
<message>Shell action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<end name="end"/>
</workflow-app>
脚本:
#!/usr/bin/env python2
import sys, os, commands, datetime, time ,getpass, errno
from optparse import OptionParser
import subprocess
from subprocess import Popen, PIPE
def arg_handle():
usage = "usage: run_validations.py [options]"
parser = OptionParser(usage)
parser.add_option("-u", "--jdbcURL",dest="jdbcURL",help="jdbcURL")
parser.add_option("-p", "--jdbcPrincipal",dest="jdbcPrincipal",help="jdbcPrincipal")
(options, args) = parser.parse_args()
print("run_validations.py -> Input : " + str(options))
return options
def main():
print("run_validations.py -> Started Run_validations.py")
options = arg_handle()
print("JDBC URL : "+options.jdbcURL)
print("JDBC PRINCIPAL : "+options.jdbcPrincipal)
beeline_connection = options.jdbcURL+";principal="+options.jdbcPrincipal
hive_cmd = 'beeline -u "'+beeline_connection+'" -e "select 1+2;"'
print("Invoked :"+hive_cmd)
rc,out = commands.getstatusoutput(hive_cmd)
if(rc==0):
print("RC : "+str(rc))
print("Output :")
print(out)
else:
print("RC : "+str(rc))
print("Output :")
print(out)
if __name__ == "__main__":
main()
输出
>>> Invoking Shell command line now >>
Stdoutput run_validations.py -> Started Run_validations.py
Stdoutput run_validations.py -> Input : {'jdbcURL': 'jdbc:hive2://host:2181/;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=hiveserver2', 'jdbcPrincipal': 'hive/_HOST@SOLON.PRD'}
Stdoutput Invoked :beeline -u "jdbc:hive2://host:2181/;serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=hiveserver2;principal=hive/_HOST@SOLON.PRD" -e "select 1+2;"
Stdoutput RC : 32512
Stdoutput Output :
Stdoutput sh: beeline: command not found
Exit code of the Shell command 0
<<< Invocation of Shell command completed <<<
有人能告诉我我错过了什么吗?你知道吗
Oozie在Hadoop集群中的另一个节点(可能是其中的一个数据节点)执行shell操作,而不是在边缘节点(测试beeline或python脚本的地方)执行shell操作。在边缘节点必须安装直线,这就是为什么你能够测试它。你知道吗
但实际问题是执行shell操作的节点似乎没有直线安装。如果您有权访问该节点,则可以登录并检查beeline。你知道吗
我建议你尝试组合蜂巢动作和壳动作来完成你想要完成的任务。你知道吗
参考:Oozie > what is the difference between asynchronous actions and synchronous actions
相关问题 更多 >
编程相关推荐