Ejemplo n.º 1
0
  def get_task_log(self, offset=0):
    logs = []
    attempt = self.task.job.job_attempts['jobAttempt'][-1]
    log_link = attempt['logsLink']
    # Get MR task logs
    if self.assignedContainerId:
      log_link = log_link.replace(attempt['containerId'], self.assignedContainerId)
    if hasattr(self, 'nodeHttpAddress'):
      log_link = log_link.replace(attempt['nodeHttpAddress'].split(':')[0], self.nodeHttpAddress.split(':')[0])

    for name in ('stdout', 'stderr', 'syslog'):
      link = '/%s/' % name
      params = {}
      if int(offset) >= 0:
        params['start'] = offset

      try:
        log_link = re.sub('job_[^/]+', self.id, log_link)
        root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
        response = root.get(link, params=params)
        log = html.fromstring(response).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
      except Exception, e:
        log = _('Failed to retrieve log: %s') % e

      logs.append(log)
Ejemplo n.º 2
0
    def get_task_log(self, offset=0):
        logs = []
        attempt = self.task.job.job_attempts["jobAttempt"][-1]
        log_link = attempt["logsLink"]
        # Get MR task logs
        if self.assignedContainerId:
            log_link = log_link.replace(attempt["containerId"], self.assignedContainerId)
        if hasattr(self, "nodeHttpAddress"):
            log_link = log_link.replace(attempt["nodeHttpAddress"].split(":")[0], self.nodeHttpAddress.split(":")[0])

        for name in ("stdout", "stderr", "syslog"):
            link = "/%s/" % name
            params = {}
            if int(offset) >= 0:
                params["start"] = offset

            try:
                log_link = re.sub("job_[^/]+", self.id, log_link)
                root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
                response = root.get(link, params=params)
                log = html.fromstring(response).xpath("/html/body/table/tbody/tr/td[2]")[0].text_content()
            except Exception, e:
                log = _("Failed to retrieve log: %s") % e

            logs.append(log)
Ejemplo n.º 3
0
def test_get_log_client():
  old_max_heap_size = clients.MAX_HEAP_SIZE
  clients.MAX_HEAP_SIZE = 2
  try:
    log_link1 = "http://test1:8041/container/nonsense"
    log_link2 = "http://test2:8041/container/nonsense"
    log_link3 = "http://test3:8041/container/nonsense"

    c1 = clients.get_log_client(log_link1)
    c2 = clients.get_log_client(log_link2)

    assert_not_equal(c1, c2)
    assert_equal(c1, clients.get_log_client(log_link1))

    clients.get_log_client(log_link3)

    assert_equal(2, len(clients._log_client_heap))
    base_urls = [tup[1].base_url for tup in clients._log_client_heap]
    assert_true('http://test1:8041' in base_urls)
    assert_true('http://test3:8041' in base_urls)
  finally:
    clients.MAX_HEAP_SIZE = old_max_heap_size
Ejemplo n.º 4
0
 def get_log_list(self):
   log_link, user = self.get_log_link()
   if not log_link:
     return []
   params = {
     'doAs': user
   }
   log_link = re.sub('job_[^/]+', str(self.id), log_link)
   root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
   response = root.get('/', params=params)
   links = html.fromstring(response, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]//a/@href')
   parsed_links = map(lambda x: urlparse.urlsplit(x), links)
   return map(lambda x: x and len(x) >= 2 and x[2].split('/')[-2] or '', parsed_links)
Ejemplo n.º 5
0
  def get_task_log(self, offset=0):
    logs = []

    log_link, user = self.get_log_link()
    if not log_link:
      return ['', '', '']

    for name in ('stdout', 'stderr', 'syslog'):
      link = '/%s/' % name
      if self.type == 'Oozie Launcher' and not self.task.job.status == 'FINISHED': # Yarn currently dumps with 500 error with doas in running state
        params = {}
      else:
        params = {
          'doAs': user
        }

      if int(offset) != 0:
        params['start'] = offset
      else:
        params['start'] = 0

      response = None
      try:
        log_link = re.sub('job_[^/]+', str(self.id), log_link)
        root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
        response = root.get(link, params=params)
        log = html.fromstring(response, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
      except Exception, e:
        log = _('Failed to retrieve log: %s' % e)
        try:
          debug_info = '\nLog Link: %s' % log_link
          if response:
            debug_info += '\nHTML Response: %s' % response
          LOG.error(debug_info)
        except:
          LOG.exception('failed to build debug info')

      logs.append(log)
Ejemplo n.º 6
0
  def get_task_log(self, offset=0):
    logs = []
    attempt = self.task.job.job_attempts['jobAttempt'][-1]
    log_link = attempt['logsLink']
    # Get MR task logs

    # Don't hack up the urls if they've been migrated to the job history server.
    for cluster in YARN_CLUSTERS.get().itervalues():
      if log_link.startswith(cluster.HISTORY_SERVER_API_URL.get()):
        break
    else:
      if self.assignedContainerId:
        log_link = log_link.replace(attempt['containerId'], self.assignedContainerId)
      if hasattr(self, 'nodeHttpAddress'):
        log_link = log_link.replace(attempt['nodeHttpAddress'].split(':')[0], self.nodeHttpAddress.split(':')[0])

    for name in ('stdout', 'stderr', 'syslog'):
      link = '/%s/' % name
      params = {}
      if int(offset) >= 0:
        params['start'] = offset

      try:
        log_link = re.sub('job_[^/]+', self.id, log_link)
        root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
        response = root.get(link, params=params)
        log = html.fromstring(response, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
      except Exception, e:
        log = _('Failed to retrieve log: %s' % e)
        try:
          debug_info = '\nLog Link: %s' % log_link
          debug_info += '\nHTML Response: %s' % response
          LOGGER.error(debug_info)
        except:
          LOG.exception('failed to build debug info')

      logs.append(log)
Ejemplo n.º 7
0
Archivo: views.py Proyecto: ronwxy/hue
def job_attempt_logs_json(request, job, attempt_index=0, name='syslog', offset=0):
  """For async log retrieval as Yarn servers are very slow"""

  try:
    attempt_index = int(attempt_index)
    attempt = job.job_attempts['jobAttempt'][attempt_index]
    log_link = attempt['logsLink']
  except (KeyError, RestException), e:
    raise KeyError(_("Cannot find job attempt '%(id)s'.") % {'id': job.jobId}, e)

  link = '/%s/' % name
  params = {}
  if offset and int(offset) >= 0:
    params['start'] = offset

  root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
  debug_info = ''
  try:
    response = root.get(link, params=params)
    log = html.fromstring(response, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
  except Exception, e:
    log = _('Failed to retrieve log: %s' % e)
    try:
      debug_info = '\nLog Link: %s' % log_link
      debug_info += '\nHTML Response: %s' % response
      LOGGER.error(debug_info)
    except:
      LOGGER.exception('failed to create debug info')

  response = {'log': LinkJobLogs._make_hdfs_links(log), 'debug': debug_info}
Ejemplo n.º 8
0
  def get_task_log(self, offset=0):
    logs = []
    attempt = self.task.job.job_attempts['jobAttempt'][-1]
    log_link = attempt['logsLink']

    # Generate actual task log link from logsLink url
    if self.task.job.status in ('NEW', 'SUBMITTED', 'RUNNING'):
      logs_path = '/node/containerlogs/'
      node_url, tracking_path = log_link.split(logs_path)
      container_id, user = tracking_path.strip('/').split('/')

      # Replace log path tokens with actual container properties if available
      if hasattr(self, 'nodeHttpAddress') and 'nodeId' in attempt:
        node_url = '%s://%s' % (node_url.split('://')[0], self.nodeHttpAddress)
      container_id = self.assignedContainerId if hasattr(self, 'assignedContainerId') else container_id

      log_link = '%(node_url)s/%(logs_path)s/%(container)s/%(user)s' % {
        'node_url': node_url,
        'logs_path': logs_path.strip('/'),
        'container': container_id,
        'user': user
      }
    else:  # Completed jobs
      logs_path = '/jobhistory/logs/'
      root_url, tracking_path = log_link.split(logs_path)
      node_url, container_id, attempt_id, user = tracking_path.strip('/').split('/')

      # Replace log path tokens with actual attempt properties if available
      if hasattr(self, 'nodeHttpAddress') and 'nodeId' in attempt:
        node_url = '%s:%s' % (self.nodeHttpAddress.split(':')[0], attempt['nodeId'].split(':')[1])
      container_id = self.assignedContainerId if hasattr(self, 'assignedContainerId') else container_id
      attempt_id = self.attemptId if hasattr(self, 'attemptId') else attempt_id

      log_link = '%(root_url)s/%(logs_path)s/%(node)s/%(container)s/%(attempt)s/%(user)s' % {
        'root_url': root_url,
        'logs_path': logs_path.strip('/'),
        'node': node_url,
        'container': container_id,
        'attempt': attempt_id,
        'user': user
      }

    for name in ('stdout', 'stderr', 'syslog'):
      link = '/%s/' % name
      params = {}
      if int(offset) != 0:
        params['start'] = offset

      response = None
      try:
        log_link = re.sub('job_[^/]+', self.id, log_link)
        root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
        response = root.get(link, params=params)
        log = html.fromstring(response, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
      except Exception, e:
        log = _('Failed to retrieve log: %s' % e)
        try:
          debug_info = '\nLog Link: %s' % log_link
          if response:
            debug_info += '\nHTML Response: %s' % response
          LOG.error(debug_info)
        except:
          LOG.exception('failed to build debug info')

      logs.append(log)
Ejemplo n.º 9
0
def job_attempt_logs_json(request,
                          job,
                          attempt_index=0,
                          name='syslog',
                          offset=LOG_OFFSET_BYTES,
                          is_embeddable=False):
    """For async log retrieval as Yarn servers are very slow"""
    log_link = None
    response = {'status': -1}

    try:
        jt = get_api(request.user, request.jt)
        app = jt.get_application(job.jobId)

        if app['applicationType'] == 'MAPREDUCE':
            if app['finalStatus'] in ('SUCCEEDED', 'FAILED', 'KILLED'):
                attempt_index = int(attempt_index)
                if not job.job_attempts['jobAttempt']:
                    response = {'status': 0, 'log': _('Job has no tasks')}
                else:
                    attempt = job.job_attempts['jobAttempt'][attempt_index]

                    log_link = attempt['logsLink']
                    # Reformat log link to use YARN RM, replace node addr with node ID addr
                    log_link = log_link.replace(attempt['nodeHttpAddress'],
                                                attempt['nodeId'])
            elif app['state'] == 'RUNNING':
                log_link = app['amContainerLogs']
        elif app.get('amContainerLogs'):
            log_link = app.get('amContainerLogs')
    except (KeyError, RestException) as e:
        raise KeyError(
            _("Cannot find job attempt '%(id)s'.") % {'id': job.jobId}, e)
    except Exception as e:
        raise Exception(
            _("Failed to get application for job %s: %s") % (job.jobId, e))

    if log_link:
        link = '/%s/' % name
        params = {'doAs': request.user.username}

        if offset != 0:
            params['start'] = offset

        root = Resource(get_log_client(log_link),
                        urllib.parse.urlsplit(log_link)[2],
                        urlencode=False)
        api_resp = None

        try:
            api_resp = root.get(link, params=params)
            log = html.fromstring(api_resp, parser=html.HTMLParser()).xpath(
                '/html/body/table/tbody/tr/td[2]')[0].text_content()

            response['status'] = 0
            response['log'] = LinkJobLogs._make_hdfs_links(log, is_embeddable)
        except Exception as e:
            response['log'] = _('Failed to retrieve log: %s' % e)
            try:
                debug_info = '\nLog Link: %s' % log_link
                if api_resp:
                    debug_info += '\nHTML Response: %s' % response
                response['debug'] = debug_info
                LOG.error(debug_info)
            except:
                LOG.exception('failed to create debug info')

    return JsonResponse(response)
Ejemplo n.º 10
0
            elif app['state'] == 'RUNNING':
                log_link = app['amContainerLogs']
    except (KeyError, RestException), e:
        raise KeyError(
            _("Cannot find job attempt '%(id)s'.") % {'id': job.jobId}, e)
    except Exception, e:
        raise Exception(
            _("Failed to get application for job %s: %s") % (job.jobId, e))

    if log_link:
        link = '/%s/' % name
        params = {'doAs': request.user.username}
        if offset != 0:
            params['start'] = offset

        root = Resource(get_log_client(log_link),
                        urlparse.urlsplit(log_link)[2],
                        urlencode=False)
        api_resp = None

        try:
            api_resp = root.get(link, params=params)
            log = html.fromstring(api_resp, parser=html.HTMLParser()).xpath(
                '/html/body/table/tbody/tr/td[2]')[0].text_content()

            response['status'] = 0
            response['log'] = LinkJobLogs._make_hdfs_links(log, is_embeddable)
        except Exception, e:
            response['log'] = _('Failed to retrieve log: %s' % e)
            try:
                debug_info = '\nLog Link: %s' % log_link
Ejemplo n.º 11
0
        # Reformat log link to use YARN RM, replace node addr with node ID addr
        log_link = log_link.replace(attempt['nodeHttpAddress'], attempt['nodeId'])
      elif app['state'] == 'RUNNING':
        log_link = app['amContainerLogs']
  except (KeyError, RestException), e:
    raise KeyError(_("Cannot find job attempt '%(id)s'.") % {'id': job.jobId}, e)
  except Exception, e:
    raise Exception(_("Failed to get application for job %s: %s") % (job.jobId, e))

  if log_link:
    link = '/%s/' % name
    params = {}
    if offset != 0:
      params['start'] = offset

    root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
    api_resp = None

    try:
      api_resp = root.get(link, params=params)
      log = html.fromstring(api_resp, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()

      response['status'] = 0
      response['log'] = LinkJobLogs._make_hdfs_links(log)
    except Exception, e:
      response['log'] = _('Failed to retrieve log: %s' % e)
      try:
        debug_info = '\nLog Link: %s' % log_link
        if api_resp:
          debug_info += '\nHTML Response: %s' % response
        response['debug'] = debug_info
Ejemplo n.º 12
0
    def get_task_log(self, offset=0):
        logs = []
        attempt = self.task.job.job_attempts['jobAttempt'][-1]
        log_link = attempt['logsLink']

        # Generate actual task log link from logsLink url
        if self.task.job.status in ('NEW', 'SUBMITTED', 'RUNNING'):
            logs_path = '/node/containerlogs/'
            node_url, tracking_path = log_link.split(logs_path)
            container_id, user = tracking_path.strip('/').split('/')

            # Replace log path tokens with actual container properties if available
            if hasattr(self, 'nodeHttpAddress') and 'nodeId' in attempt:
                node_url = '%s://%s:%s' % (node_url.split('://')[0],
                                           self.nodeHttpAddress.split(':')[0],
                                           attempt['nodeId'].split(':')[1])
            container_id = self.assignedContainerId if hasattr(
                self, 'assignedContainerId') else container_id

            log_link = '%(node_url)s/%(logs_path)s/%(container)s/%(user)s' % {
                'node_url': node_url,
                'logs_path': logs_path.strip('/'),
                'container': container_id,
                'user': user
            }
        else:  # Completed jobs
            logs_path = '/jobhistory/logs/'
            root_url, tracking_path = log_link.split(logs_path)
            node_url, container_id, attempt_id, user = tracking_path.strip(
                '/').split('/')

            # Replace log path tokens with actual attempt properties if available
            if hasattr(self, 'nodeHttpAddress') and 'nodeId' in attempt:
                node_url = '%s:%s' % (self.nodeHttpAddress.split(':')[0],
                                      attempt['nodeId'].split(':')[1])
            container_id = self.assignedContainerId if hasattr(
                self, 'assignedContainerId') else container_id
            attempt_id = self.attemptId if hasattr(self,
                                                   'attemptId') else attempt_id

            log_link = '%(root_url)s/%(logs_path)s/%(node)s/%(container)s/%(attempt)s/%(user)s' % {
                'root_url': root_url,
                'logs_path': logs_path.strip('/'),
                'node': node_url,
                'container': container_id,
                'attempt': attempt_id,
                'user': user
            }

        for name in ('stdout', 'stderr', 'syslog'):
            link = '/%s/' % name
            params = {}
            if int(offset) >= 0:
                params['start'] = offset

            try:
                log_link = re.sub('job_[^/]+', self.id, log_link)
                root = Resource(get_log_client(log_link),
                                urlparse.urlsplit(log_link)[2],
                                urlencode=False)
                response = root.get(link, params=params)
                log = html.fromstring(
                    response, parser=html.HTMLParser()).xpath(
                        '/html/body/table/tbody/tr/td[2]')[0].text_content()
            except Exception, e:
                log = _('Failed to retrieve log: %s' % e)
                try:
                    debug_info = '\nLog Link: %s' % log_link
                    debug_info += '\nHTML Response: %s' % response
                    LOG.error(debug_info)
                except:
                    LOG.exception('failed to build debug info')

            logs.append(log)
Ejemplo n.º 13
0
def job_attempt_logs_json(request, job, attempt_index=0, name='syslog', offset=0):
  """For async log retrieval as Yarn servers are very slow"""

  try:
    attempt_index = int(attempt_index)
    attempt = job.job_attempts['jobAttempt'][attempt_index]
    log_link = attempt['logsLink']
  except (KeyError, RestException), e:
    raise KeyError(_("Cannot find job attempt '%(id)s'.") % {'id': job.jobId}, e)

  link = '/%s/' % name
  params = {}
  if offset and int(offset) >= 0:
    params['start'] = offset

  root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2])

  try:
    response = root.get(link, params=params)
    log = html.fromstring(response).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
  except Exception, e:
    log = _('Failed to retrieve log: %s') % e

  response = {'log': log}

  return HttpResponse(json.dumps(response), mimetype="application/json")



@check_job_permission
def job_single_logs(request, job):