Merge all swarm_*.py scripts into swarming.py.

swarming.py

+#!/usr/bin/env python
+# Copyright (c) 2012 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import hashlib
+import json
+import logging
+import os
+import re
+import shutil
+import StringIO
+import subprocess
+import sys
+import threading
+import time
+import urllib
+import zipfile
+
+from third_party.depot_tools import fix_encoding
+from third_party.depot_tools import subcommand
+
+import run_isolated
+import trace_inputs
+
+
+ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
+TOOLS_PATH = os.path.join(ROOT_DIR, 'tools')
+
+
+# Default servers.
+# TODO(maruel): Chromium-specific.
+ISOLATE_SERVER = 'https://isolateserver-dev.appspot.com/'
+SWARM_SERVER = 'https://chromium-swarm-dev.appspot.com'
+
+
+# The default time to wait for a shard to finish running.
+DEFAULT_SHARD_WAIT_TIME = 40 * 60.
+
+
+PLATFORM_MAPPING = {
+  'cygwin': 'Windows',
+  'darwin': 'Mac',
+  'linux2': 'Linux',
+  'win32': 'Windows',
+}
+
+
+class Failure(Exception):
+  """Generic failure."""
+  pass
+
+
+class Manifest(object):
+  """Represents a Swarming task manifest.
+
+  Also includes code to zip code and upload itself.
+  """
+  def __init__(
+      self, manifest_hash, test_name, shards, test_filter, os_image,
+      working_dir, cac, verbose, profile, priority):
+    """Populates a manifest object.
+      Args:
+        manifest_hash - The manifest's sha-1 that the slave is going to fetch.
+        test_name - The name to give the test request.
+        shards - The number of swarm shards to request.
+        test_filter - The gtest filter to apply when running the test.
+        os_image - OS to run on.
+        working_dir - Relative working directory to start the script.
+        cac - isolate server url.
+        verbose - if True, have the slave print more details.
+        profile - if True, have the slave print more timing data.
+        priority - int between 0 and 1000, lower the higher priority
+    """
+    self.manifest_hash = manifest_hash
+    self._test_name = test_name
+    self._shards = shards
+    self._test_filter = test_filter
+    self._target_platform = PLATFORM_MAPPING[os_image]
+    self._working_dir = working_dir
+
+    base_url = cac.rstrip('/')
+    self.data_server_retrieval = base_url + '/content/retrieve/default/'
+    self._data_server_storage = base_url + '/content/store/default/'
+    self._data_server_has = base_url + '/content/contains/default'
+    self._data_server_get_token = base_url + '/content/get_token'
+
+    self.verbose = bool(verbose)
+    self.profile = bool(profile)
+    self.priority = priority
+
+    self._zip_file_hash = ''
+    self._tasks = []
+    self._files = {}
+    self._token_cache = None
+
+  def _token(self):
+    if not self._token_cache:
+      result = run_isolated.url_open(self._data_server_get_token)
+      if not result:
+        # TODO(maruel): Implement authentication.
+        raise Failure('Failed to get token, need authentication')
+      # Quote it right away, so creating the urls is simpler.
+      self._token_cache = urllib.quote(result.read())
+    return self._token_cache
+
+  def add_task(self, task_name, actions, time_out=600):
+    """Appends a new task to the swarm manifest file."""
+    # See swarming/src/common/test_request_message.py TestObject constructor for
+    # the valid flags.
+    self._tasks.append(
+        {
+          'action': actions,
+          'decorate_output': self.verbose,
+          'test_name': task_name,
+          'time_out': time_out,
+        })
+
+  def add_file(self, source_path, rel_path):
+    self._files[source_path] = rel_path
+
+  def zip_and_upload(self):
+    """Zips up all the files necessary to run a shard and uploads to Swarming
+    master.
+    """
+    assert not self._zip_file_hash
+    start_time = time.time()
+
+    zip_memory_file = StringIO.StringIO()
+    zip_file = zipfile.ZipFile(zip_memory_file, 'w')
+
+    for source, relpath in self._files.iteritems():
+      zip_file.write(source, relpath)
+
+    zip_file.close()
+    print 'Zipping completed, time elapsed: %f' % (time.time() - start_time)
+
+    zip_memory_file.flush()
+    zip_contents = zip_memory_file.getvalue()
+    zip_memory_file.close()
+
+    self._zip_file_hash = hashlib.sha1(zip_contents).hexdigest()
+
+    response = run_isolated.url_open(
+        self._data_server_has + '?token=%s' % self._token(),
+        data=self._zip_file_hash,
+        content_type='application/octet-stream')
+    if response is None:
+      print >> sys.stderr, (
+          'Unable to query server for zip file presence, aborting.')
+      return False
+
+    if response.read(1) == chr(1):
+      print 'Zip file already on server, no need to reupload.'
+      return True
+
+    print 'Zip file not on server, starting uploading.'
+
+    url = '%s%s?priority=0&token=%s' % (
+        self._data_server_storage, self._zip_file_hash, self._token())
+    response = run_isolated.url_open(
+        url, data=zip_contents, content_type='application/octet-stream')
+    if response is None:
+      print >> sys.stderr, 'Failed to upload the zip file: %s' % url
+      return False
+
+    return True
+
+  def to_json(self):
+    """Exports the current configuration into a swarm-readable manifest file.
+
+    This function doesn't mutate the object.
+    """
+    test_case = {
+      'test_case_name': self._test_name,
+      'data': [
+        [self.data_server_retrieval + urllib.quote(self._zip_file_hash),
+         'swarm_data.zip'],
+      ],
+      'tests': self._tasks,
+      'env_vars': {},
+      'configurations': [
+        {
+          'min_instances': self._shards,
+          'config_name': self._target_platform,
+          'dimensions': {
+            'os': self._target_platform,
+          },
+        },
+      ],
+      'working_dir': self._working_dir,
+      'restart_on_failure': True,
+      'cleanup': 'root',
+      'priority': self.priority,
+    }
+
+    # These flags are googletest specific.
+    if self._test_filter and self._test_filter != '*':
+      test_case['env_vars']['GTEST_FILTER'] = self._test_filter
+    if self._shards > 1:
+      test_case['env_vars']['GTEST_SHARD_INDEX'] = '%(instance_index)s'
+      test_case['env_vars']['GTEST_TOTAL_SHARDS'] = '%(num_instances)s'
+
+    return json.dumps(test_case, separators=(',',':'))
+
+
+class Bit(object):
+  """Thread safe setable bit."""
+  _lock = threading.Lock()
+  _value = False
+
+  def get(self):
+    with self._lock:
+      return self._value
+
+  def set(self):
+    with self._lock:
+      self._value = True
+
+
+def now():
+  """Exists so it can be mocked easily."""
+  return time.time()
+
+
+def get_test_keys(swarm_base_url, test_name):
+  """Returns the Swarm test key for each shards of test_name."""
+  key_data = urllib.urlencode([('name', test_name)])
+  url = '%s/get_matching_test_cases?%s' % (swarm_base_url, key_data)
+
+  for i in range(run_isolated.URL_OPEN_MAX_ATTEMPTS):
+    response = run_isolated.url_open(url, retry_404=True)
+    if response is None:
+      raise Failure(
+          'Error: Unable to find any tests with the name, %s, on swarm server'
+          % test_name)
+
+    result = response.read()
+    # TODO(maruel): Compare exact string.
+    if 'No matching' in result:
+      logging.warning('Unable to find any tests with the name, %s, on swarm '
+                      'server' % test_name)
+      if i != run_isolated.URL_OPEN_MAX_ATTEMPTS:
+        run_isolated.HttpService.sleep_before_retry(i, None)
+      continue
+    return json.loads(result)
+
+  raise Failure(
+      'Error: Unable to find any tests with the name, %s, on swarm server'
+      % test_name)
+
+
+def retrieve_results(base_url, test_key, timeout, should_stop):
+  """Retrieves results for a single test_key."""
+  assert isinstance(timeout, float)
+  params = [('r', test_key)]
+  result_url = '%s/get_result?%s' % (base_url, urllib.urlencode(params))
+  start = now()
+  while True:
+    if timeout and (now() - start) >= timeout:
+      logging.error('retrieve_results(%s) timed out', base_url)
+      return {}
+    # Do retries ourselves.
+    response = run_isolated.url_open(
+        result_url, retry_404=False, retry_50x=False)
+    if response is None:
+      # Aggressively poll for results. Do not use retry_404 so
+      # should_stop is polled more often.
+      remaining = min(5, timeout - (now() - start)) if timeout else 5
+      if remaining > 0:
+        run_isolated.HttpService.sleep_before_retry(1, remaining)
+    else:
+      try:
+        data = json.load(response) or {}
+      except (ValueError, TypeError):
+        logging.warning(
+            'Received corrupted data for test_key %s. Retrying.', test_key)
+      else:
+        if data['output']:
+          return data
+    if should_stop.get():
+      return {}
+
+
+def yield_results(swarm_base_url, test_keys, timeout, max_threads):
+  """Yields swarm test results from the swarm server as (index, result).
+
+  Duplicate shards are ignored, the first one to complete is returned.
+
+  max_threads is optional and is used to limit the number of parallel fetches
+  done. Since in general the number of test_keys is in the range <=10, it's not
+  worth normally to limit the number threads. Mostly used for testing purposes.
+  """
+  shards_remaining = range(len(test_keys))
+  number_threads = (
+      min(max_threads, len(test_keys)) if max_threads else len(test_keys))
+  should_stop = Bit()
+  results_remaining = len(test_keys)
+  with run_isolated.ThreadPool(number_threads, number_threads, 0) as pool:
+    try:
+      for test_key in test_keys:
+        pool.add_task(
+            0, retrieve_results, swarm_base_url, test_key, timeout, should_stop)
+      while shards_remaining and results_remaining:
+        result = pool.get_one_result()
+        results_remaining -= 1
+        if not result:
+          # Failed to retrieve one key.
+          logging.error('Failed to retrieve the results for a swarm key')
+          continue
+        shard_index = result['config_instance_index']
+        if shard_index in shards_remaining:
+          shards_remaining.remove(shard_index)
+          yield shard_index, result
+        else:
+          logging.warning('Ignoring duplicate shard index %d', shard_index)
+          # Pop the last entry, there's no such shard.
+          shards_remaining.pop()
+    finally:
+      # Done, kill the remaining threads.
+      should_stop.set()
+
+
+def chromium_setup(manifest):
+  """Sets up the commands to run.
+
+  Highly chromium specific.
+  """
+  cleanup_script_name = 'swarm_cleanup.py'
+  cleanup_script_path = os.path.join(TOOLS_PATH, cleanup_script_name)
+  run_test_name = 'run_isolated.py'
+  run_test_path = os.path.join(ROOT_DIR, run_test_name)
+
+  manifest.add_file(run_test_path, run_test_name)
+  manifest.add_file(cleanup_script_path, cleanup_script_name)
+  run_cmd = [
+    'python', run_test_name,
+    '--hash', manifest.manifest_hash,
+    '--remote', manifest.data_server_retrieval.rstrip('/') + '-gzip/',
+  ]
+  if manifest.verbose or manifest.profile:
+    # Have it print the profiling section.
+    run_cmd.append('--verbose')
+  manifest.add_task('Run Test', run_cmd)
+
+  # Clean up
+  manifest.add_task('Clean Up', ['python', cleanup_script_name])
+
+
+def archive(isolated, cac, verbose):
+  """Archives a .isolated and all the dependencies on the CAC."""
+  tempdir = None
+  try:
+    logging.info('Archiving')
+    cmd = [
+      'isolate.py',
+      'hashtable',
+      '--outdir', cac,
+      '--isolated', isolated,
+    ]
+    if verbose:
+      cmd.append('--verbose')
+    if subprocess.call(cmd, verbose):
+      return
+    return hashlib.sha1(open(isolated, 'rb').read()).hexdigest()
+  finally:
+    if tempdir:
+      shutil.rmtree(tempdir)
+
+
+def process_manifest(
+    file_sha1_or_isolated, test_name, shards, test_filter, os_image,
+    working_dir, cac, swarm_url, verbose, profile, priority):
+  """Process the manifest file and send off the swarm test request.
+
+  Optionally archives an .isolated file.
+  """
+  if file_sha1_or_isolated.endswith('.isolated'):
+    file_sha1 = archive(file_sha1_or_isolated, swarm_url, verbose)
+    if not file_sha1:
+      print >> sys.stderr, 'Archival failure %s' % file_sha1_or_isolated
+      return 1
+  elif re.match(r'^[a-f0-9]{40}$', file_sha1_or_isolated):
+    file_sha1 = file_sha1_or_isolated
+  else:
+    print >> sys.stderr, 'Invalid hash %s' % file_sha1_or_isolated
+    return 1
+
+  try:
+    manifest = Manifest(
+        file_sha1, test_name, shards, test_filter, os_image,
+        working_dir, cac, verbose, profile, priority)
+  except ValueError as e:
+    print >> sys.stderr, 'Unable to process %s: %s' % (test_name, e)
+    return 1
+
+  chromium_setup(manifest)
+
+  # Zip up relevant files.
+  print('Zipping up files...')
+  if not manifest.zip_and_upload():
+    return 1
+
+  # Send test requests off to swarm.
+  print('Sending test requests to swarm.')
+  print('Server: %s' % swarm_url)
+  print('Job name: %s' % test_name)
+  test_url = swarm_url.rstrip('/') + '/test'
+  manifest_text = manifest.to_json()
+  result = run_isolated.url_open(test_url, data={'request': manifest_text})
+  if not result:
+    print >> sys.stderr, 'Failed to send test for %s\n%s' % (
+        test_name, test_url)
+    return 1
+  try:
+    json.load(result)
+  except (ValueError, TypeError) as e:
+    print >> sys.stderr, 'Failed to send test for %s' % test_name
+    print >> sys.stderr, 'Manifest: %s' % manifest_text
+    print >> sys.stderr, str(e)
+    return 1
+  return 0
+
+
+def trigger(
+    os_image,

[Vadim Sh., 2013/08/19 18:55:44]

That's one hell of an argument list.

Maybe put some of this stuff into utility object and pass it around?

Stuff than can go there: os_image, test_name_prefix, working_dir, cac,
swarm_url, verbose, profile, priority (so it can be passed as is to
process_manifest).

Alternatively 'trigger' and 'process_manifest' can be converted to a methods of
some 'SwarmTrigger' class that can hold all this details.

[M-A Ruel, 2013/08/20 17:09:30]

The Manifest object will have to be redesigned, but I prefer to punt that on a
follow up CL with no user impacting change.

+    tasks,
+    test_name_prefix,
+    working_dir,
+    cac,
+    swarm_url,
+    verbose,
+    profile,
+    priority):
+  """Sends off the hash swarming test requests."""
+  if not os_image or os_image == 'None':
+    # Use the current OS.
+    os_image = sys.platform
+
+  highest_exit_code = 0
+  for (file_sha1, test_name, shards, testfilter) in tasks:
+    exit_code = process_manifest(
+        file_sha1,
+        test_name_prefix + test_name,
+        int(shards),
+        testfilter,
+        os_image,
+        working_dir,
+        cac,
+        swarm_url,
+        verbose,
+        profile,
+        priority)
+    highest_exit_code = max(highest_exit_code, exit_code)
+  return highest_exit_code
+
+
+def collect(url, test_name, timeout):
+  """Retrieves results of a Swarming job."""
+  test_keys = get_test_keys(url, test_name)
+  if not test_keys:
+    raise Failure('No test keys to get results with.')
+
+  exit_code = 0
+  for _index, output in yield_results(url, test_keys, timeout, None):
+    print(
+        '%s/%s: %s' % (
+            output['machine_id'], output['machine_tag'], output['exit_codes']))
+    print(''.join('  %s\n' % l for l in output['output'].splitlines()))
+    exit_code = max(exit_code, max(map(int, output['exit_codes'].split(','))))
+  return exit_code
+
+
+def add_trigger_options(parser):
+  """Adds all options to trigger a task on Swarming."""
+  parser.add_option(
+      '-c', '--cac',

[M-A Ruel, 2013/08/18 00:18:19]

I don't like --data_server. --isolate_server (like I still have left in
run_example_swarm.py) is longish and I'd like to rename the server, since it's
not that isolate-specific.

isolate.py's --outdir is just wrong but that's for another time.
run_isolated.py's --remote is weird.

Any suggestion? They should all be changed to the same name in any case. I like
--cac but it's not clear what it means and it's not obvious it's acronym.

+      default=ISOLATE_SERVER,
+      metavar='URL',
+      help='Isolate server where data is stored. default: %default')
+  parser.add_option(
+      '-w', '--working_dir', default='swarm_tests',
+      help='Working directory on the swarm slave side. default: %default.')
+  parser.add_option(
+      '-o', '--os_image',

[M-A Ruel, 2013/08/18 00:18:19]

--os makes sense, but --os_image is weird.

[Vadim Sh., 2013/08/19 18:55:44]

If it's 'one of valid sys.platform' maybe call it 'platform'? :)

[M-A Ruel, 2013/08/20 17:09:30]

platform is more ambiguous than 'os'

+      help='Swarm OS image to request. Should be one of the valid sys.platform '
+           'values like darwin, linux2 or win32.')
+  parser.add_option(
+      '-T', '--test-name-prefix', default='',

[Vadim Sh., 2013/08/19 18:55:44]

I personally prefer dashes in argument names (like here). So I propose to change
everything else to use dashes as well. 

If it's not possible, then change --test-name-prefix to --test_name_prefix for
consistency.

[M-A Ruel, 2013/08/20 17:09:30]

--task-prefix.

+      help='Prefix to give the swarm test request. default: %default')
+  parser.add_option(
+      '--profile', action='store_true',
+      default=bool(os.environ.get('ISOLATE_DEBUG')),
+      help='Have run_isolated.py print profiling info')
+  parser.add_option(
+      '--priority', type='int', default=100,
+      help='The lower value, the more important the task is')
+
+
+def add_collect_options(parser):
+  parser.add_option(
+      '-t', '--timeout',
+      type='float',
+      default=DEFAULT_SHARD_WAIT_TIME,
+      help='Timeout to wait for result, set to 0 for no timeout; default: '
+           '%default s')
+
+
+@subcommand.usage('test_name')
+def CMDcollect(parser, args):
+  """Retrieves results of a Swarming job.
+
+  The result can be in multiple part if the execution was sharded. It can
+  potentially have retries.
+  """
+  add_collect_options(parser)
+  (options, args) = parser.parse_args(args)
+  if not args:
+    parser.error('Must specify one test name.')
+  elif len(args) > 1:
+    parser.error('Must specify only one test name.')
+
+  try:
+    return collect(options.server, args[0], options.timeout)
+  except Failure as e:
+    parser.error(e.args[0])
+
+
+@subcommand.usage('[sha1|isolated ...]')
+def CMDrun(parser, args):
+  """Trigger a job and wait for the results.
+
+  Basically, does everything to run a command remotely.
+  """
+  add_trigger_options(parser)
+  add_collect_options(parser)
+  options, args = parser.parse_args(args)
+
+  if not args:
+    parser.error('Must pass at least one .isolated file or its sha1.')
+
+  success = []
+  for arg in args:
+    try:
+      result = trigger(
+          options.os_image,
+          options.tasks,
+          options.test_name_prefix,
+          options.working_dir,
+          options.cac,
+          options.swarm_url,
+          options.verbose,
+          options.profile,
+          options.priority)
+    except Failure as e:
+      parser.error(e.args[0])
+    success.append(arg)
+
+  if not success:
+    return result
+
+  code = 0
+  for arg in success:
+    try:
+      code = min(code, collect(options.server, arg, options.timeout))
+    except Failure as e:
+      result = 1
+      print >> sys.stderr, e.args[0]
+  return code
+
+
+def CMDtrigger(parser, args):
+  """Triggers Swarm request(s).
+
+  Accepts one or multiple --task requests, with either the sha1 of a .isolated
+  file already uploaded or the path to an .isolated file to archive, packages it
+  if needed and sends a Swarm manifest file to the Swarm server.
+  """
+  add_trigger_options(parser)
+  parser.add_option(
+      '--task', nargs=4, action='append', default=[], dest='tasks',

[M-A Ruel, 2013/08/18 00:18:19]

I like --task contrary to --run_from_hash, especially that I changed it to
support specifying a .isolated file directly.

+      help='Task to trigger. The format is '
+           '(hash|isolated, test_name, shards, test_filter). This may be '
+           'used multiple times to send multiple hashes jobs. If an isolated '
+           'file is specified instead of an hash, it is first archived.')
+  (options, args) = parser.parse_args(args)
+
+  if args:
+    parser.error('Unknown args: %s' % args)
+  if not options.tasks:
+    parser.error('At least one --task is required.')
+  if not options.cac:
+    parser.error('Must specify the CAC server.')
+
+  try:
+    return trigger(
+        options.os_image,
+        options.tasks,
+        options.test_name_prefix,
+        options.working_dir,
+        options.cac,
+        options.server,
+        options.verbose,
+        options.profile,
+        options.priority)
+  except Failure as e:
+    parser.error(e.args[0])
+
+
+class OptionParserSwarming(trace_inputs.OptionParserWithLogging):
+  def __init__(self, **kwargs):
+    trace_inputs.OptionParserWithLogging.__init__(
+        self, prog='swarming.py', **kwargs)
+    self.add_option(
+        '-s', '--server', default=SWARM_SERVER,

[M-A Ruel, 2013/08/18 00:18:19]

--server is too generic.
--swarm_url, for an unknown reason, I dislike.
--url is bad.
--swarming is unclear.

Any suggestion?

+        help='Specify the url of the Swarming server, default: %default')
+
+  def parse_args(self, *args, **kwargs):
+    options, args = trace_inputs.OptionParserWithLogging.parse_args(
+        self, *args, **kwargs)
+    options.server = options.server.rstrip('/')
+    if not options.server:
+      self.error('--server is required')
+    return options, args
+
+
+def main(args):
+  dispatcher = subcommand.CommandDispatcher(__name__)
+  try:
+    return dispatcher.execute(OptionParserSwarming(), args)
+  except (
+      Failure,
+      run_isolated.MappingError,
+      run_isolated.ConfigError) as e:
+    sys.stderr.write('\nError: ')
+    sys.stderr.write(str(e))
+    sys.stderr.write('\n')
+    return 1
+
+
+if __name__ == '__main__':
+  fix_encoding.fix_encoding()
+  run_isolated.disable_buffering()
+  sys.exit(main(sys.argv[1:]))