pyspark 以HA方式访问hdfs

    sc, sqlContext = s.startContext(appName)
   
    sc._jsc.hadoopConfiguration().set("dfs.nameservices", "umecluster2")
    sc._jsc.hadoopConfiguration().set('dfs.ha.namenodes.umecluster2', 'nn1,nn2')
    sc._jsc.hadoopConfiguration().set('dfs.namenode.rpc-address.umecluster2.nn2', 'hdfs://10.5.145.xxx:8020')
    sc._jsc.hadoopConfiguration().set('dfs.namenode.rpc-address.umecluster2.nn1', 'hdfs://10.5.145.xxx:8020')
    sc._jsc.hadoopConfiguration().set("dfs.client.failover.proxy.provider.umecluster2", "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider")