如何用python客户端实时捕获HiveServer2的查询日志



我使用pyhs2 (https://pypi.python.org/pypi/pyhs2)的修改版本,能够运行异步查询和TCLIService的其他方法。客户端(GetLog, send_GetLog, recv_GetLog)在Hue的源代码(https://github.com/cloudera/hue/blob/master/apps/beeswax/gen-py/TCLIService/TCLIService.py#L739)

但是当我运行TCLIService.Client.GetLog方法时,有一个错误:

$ python example.py 
Traceback (most recent call last):
  File "example.py", line 85, in <module>
    rq = client.GetLog(lq)
  File "/Users/toly/hive_streaming/libs/pyhs4/TCLIService/TCLIService.py", line 757, in GetLog
    return self.recv_GetLog()
  File "/Users/toly/hive_streaming/libs/pyhs4/TCLIService/TCLIService.py", line 773, in recv_GetLog
    raise x
thrift.Thrift.TApplicationException: Invalid method name: 'GetLog'

在脚本中,我使用来自Cloudera VM的HiveServer2。我猜Hue使用的是同一台服务器,它成功地工作了。此外,我尝试了client_protocol在0到7的范围内创建会话。

import time
import sasl
from thrift.protocol.TBinaryProtocol import TBinaryProtocol
from thrift.transport.TSocket import TSocket
from thrift.transport.TTransport import TBufferedTransport
from libs.pyhs4.cloudera.thrift_sasl import TSaslClientTransport

from libs.pyhs4.TCLIService import TCLIService
from libs.pyhs4.TCLIService.ttypes import TOpenSessionReq, TGetTablesReq, TFetchResultsReq,
    TStatusCode, TGetResultSetMetadataReq, TGetColumnsReq, TType, TTypeId, 
    TExecuteStatementReq, TGetOperationStatusReq, TFetchOrientation, TCloseOperationReq, 
    TCloseSessionReq, TGetSchemasReq, TCancelOperationReq, TGetLogReq
auth = 'PLAIN'
username = 'apanin'
password = 'none'
host = 'cloudera'
port = 10000
test_hql1 = 'select count(*) from test_text'

def sasl_factory():
    saslc = sasl.Client()
    saslc.setAttr("username", username)
    saslc.setAttr("password", password)
    saslc.init()
    return saslc

def get_type(typeDesc):
    for ttype in typeDesc.types:
        if ttype.primitiveEntry is not None:
            return TTypeId._VALUES_TO_NAMES[ttype.primitiveEntry.type]
        elif ttype.mapEntry is not None:
            return ttype.mapEntry
        elif ttype.unionEntry is not None:
            return ttype.unionEntry
        elif ttype.arrayEntry is not None:
            return ttype.arrayEntry
        elif ttype.structEntry is not None:
            return ttype.structEntry
        elif ttype.userDefinedTypeEntry is not None:
            return ttype.userDefinedTypeEntry

def get_value(colValue):
    if colValue.boolVal is not None:
      return colValue.boolVal.value
    elif colValue.byteVal is not None:
      return colValue.byteVal.value
    elif colValue.i16Val is not None:
      return colValue.i16Val.value
    elif colValue.i32Val is not None:
      return colValue.i32Val.value
    elif colValue.i64Val is not None:
      return colValue.i64Val.value
    elif colValue.doubleVal is not None:
      return colValue.doubleVal.value
    elif colValue.stringVal is not None:
      return colValue.stringVal.value

sock = TSocket(host, port)
transport = TSaslClientTransport(sasl_factory, "PLAIN", sock)
client = TCLIService.Client(TBinaryProtocol(transport))
transport.open()
res = client.OpenSession(TOpenSessionReq(username=username, password=password))
session = res.sessionHandle
query1 = TExecuteStatementReq(session, statement=test_hql1, confOverlay={}, runAsync=True)
response1 = client.ExecuteStatement(query1)
opHandle1 = response1.operationHandle

while True:
    time.sleep(1)
    q1 = TGetOperationStatusReq(operationHandle=opHandle1)
    res1 = client.GetOperationStatus(q1)
    lq = TGetLogReq(opHandle1)
    rq = client.GetLog(lq)
    if res1.operationState == 2:
        break

req = TCloseOperationReq(operationHandle=opHandle1)
client.CloseOperation(req)
req = TCloseSessionReq(sessionHandle=session)
client.CloseSession(req)

如何从HiveServer2实时捕获hive查询日志?

乌利希期刊指南蜂巢版- 1.2.1 "

获取方法FetchResults的操作日志,参数为fetchType=1 -返回日志

使用例子:

query1 = TExecuteStatementReq(session, statement=test_hql1, confOverlay={}, runAsync=True)
response1 = client.ExecuteStatement(query1)
opHandle1 = response1.operationHandle
while True:
    time.sleep(1)
    q1 = TGetOperationStatusReq(operationHandle=opHandle1)
    res1 = client.GetOperationStatus(q1)
    request_logs = TFetchResultsReq(operationHandle=opHandle1, orientation=0, maxRows=10, fetchType=1)
    response_logs = client.FetchResults(request_logs)
    print response_logs.results
    if res1.operationState == 2:
        break

最新更新