def test_num_cpus():
with pytest.deprecated_call():
sysinfo.num_cpus()
class IPClusterEngines(BaseParallelApplication):
name = u'ipcluster'
description = engines_help
examples = _engines_examples
usage = None
default_log_level = logging.INFO
classes = List()
def _classes_default(self):
from IPython.parallel.apps import launcher
launchers = launcher.all_launchers
eslaunchers = [ l for l in launchers if 'EngineSet' in l.__name__]
return [ProfileDir]+eslaunchers
n = Integer(num_cpus(), config=True,
help="""The number of engines to start. The default is to use one for each
CPU on your machine""")
engine_launcher = Any(config=True, help="Deprecated, use engine_launcher_class")
def _engine_launcher_changed(self, name, old, new):
if isinstance(new, string_types):
self.log.warn("WARNING: %s.engine_launcher is deprecated as of 0.12,"
" use engine_launcher_class" % self.__class__.__name__)
self.engine_launcher_class = new
engine_launcher_class = DottedObjectName('LocalEngineSetLauncher',
config=True,
help="""The class for launching a set of Engines. Change this value
to use various batch systems to launch your engines, such as PBS,SGE,MPI,etc.
Each launcher class has its own set of configuration options, for making sure
it will work in your environment.
You can also write your own launcher, and specify it's absolute import path,
as in 'mymodule.launcher.FTLEnginesLauncher`.
IPython's bundled examples include:
Local : start engines locally as subprocesses [default]
MPI : use mpiexec to launch engines in an MPI environment
PBS : use PBS (qsub) to submit engines to a batch queue
SGE : use SGE (qsub) to submit engines to a batch queue
LSF : use LSF (bsub) to submit engines to a batch queue
SSH : use SSH to start the controller
Note that SSH does *not* move the connection files
around, so you will likely have to do this manually
unless the machines are on a shared file system.
HTCondor : use HTCondor to submit engines to a batch queue
WindowsHPC : use Windows HPC
If you are using one of IPython's builtin launchers, you can specify just the
prefix, e.g:
c.IPClusterEngines.engine_launcher_class = 'SSH'
or:
ipcluster start --engines=MPI
"""
)
daemonize = Bool(False, config=True,
help="""Daemonize the ipcluster program. This implies --log-to-file.
Not available on Windows.
""")
def _daemonize_changed(self, name, old, new):
if new:
self.log_to_file = True
early_shutdown = Integer(30, config=True, help="The timeout (in seconds)")
_stopping = False
aliases = Dict(engine_aliases)
flags = Dict(engine_flags)
@catch_config_error
def initialize(self, argv=None):
super(IPClusterEngines, self).initialize(argv)
self.init_signal()
self.init_launchers()
def init_launchers(self):
self.engine_launcher = self.build_launcher(self.engine_launcher_class, 'EngineSet')
def init_signal(self):
# Setup signals
signal.signal(signal.SIGINT, self.sigint_handler)
def build_launcher(self, clsname, kind=None):
"""import and instantiate a Launcher based on importstring"""
try:
klass = find_launcher_class(clsname, kind)
except (ImportError, KeyError):
self.log.fatal("Could not import launcher class: %r"%clsname)
self.exit(1)
launcher = klass(
work_dir=u'.', parent=self, log=self.log,
profile_dir=self.profile_dir.location, cluster_id=self.cluster_id,
)
return launcher
def engines_started_ok(self):
self.log.info("Engines appear to have started successfully")
self.early_shutdown = 0
def start_engines(self):
# Some EngineSetLaunchers ignore `n` and use their own engine count, such as SSH:
n = getattr(self.engine_launcher, 'engine_count', self.n)
self.log.info("Starting %s Engines with %s", n, self.engine_launcher_class)
try:
self.engine_launcher.start(self.n)
except:
self.log.exception("Engine start failed")
raise
self.engine_launcher.on_stop(self.engines_stopped_early)
if self.early_shutdown:
ioloop.DelayedCallback(self.engines_started_ok, self.early_shutdown*1000, self.loop).start()
def engines_stopped_early(self, r):
if self.early_shutdown and not self._stopping:
self.log.error("""
Engines shutdown early, they probably failed to connect.
Check the engine log files for output.
If your controller and engines are not on the same machine, you probably
have to instruct the controller to listen on an interface other than localhost.
You can set this by adding "--ip='*'" to your ControllerLauncher.controller_args.
Be sure to read our security docs before instructing your controller to listen on
a public interface.
""")
self.stop_launchers()
return self.engines_stopped(r)
def engines_stopped(self, r):
return self.loop.stop()
def stop_engines(self):
if self.engine_launcher.running:
self.log.info("Stopping Engines...")
d = self.engine_launcher.stop()
return d
else:
return None
def stop_launchers(self, r=None):
if not self._stopping:
self._stopping = True
self.log.error("IPython cluster: stopping")
self.stop_engines()
# Wait a few seconds to let things shut down.
dc = ioloop.DelayedCallback(self.loop.stop, 3000, self.loop)
dc.start()
def sigint_handler(self, signum, frame):
self.log.debug("SIGINT received, stopping launchers...")
self.stop_launchers()
def start_logging(self):
# Remove old log files of the controller and engine
if self.clean_logs:
log_dir = self.profile_dir.log_dir
for f in os.listdir(log_dir):
if re.match(r'ip(engine|controller)-.+\.(log|err|out)',f):
os.remove(os.path.join(log_dir, f))
def start(self):
"""Start the app for the engines subcommand."""
self.log.info("IPython cluster: started")
# First see if the cluster is already running
# Now log and daemonize
self.log.info(
'Starting engines with [daemon=%r]' % self.daemonize
)
# TODO: Get daemonize working on Windows or as a Windows Server.
if self.daemonize:
if os.name=='posix':
daemonize()
dc = ioloop.DelayedCallback(self.start_engines, 0, self.loop)
dc.start()
# Now write the new pid file AFTER our new forked pid is active.
# self.write_pid_file()
try:
self.loop.start()
except KeyboardInterrupt:
pass
except zmq.ZMQError as e:
if e.errno == errno.EINTR:
pass
else:
raise
class IPClusterEngines(BaseParallelApplication):
name = u'ipcluster'
description = engines_help
examples = _engines_examples
usage = None
config_file_name = Unicode(default_config_file_name)
default_log_level = logging.INFO
classes = List()
def _classes_default(self):
from IPython.parallel.apps import launcher
launchers = launcher.all_launchers
eslaunchers = [l for l in launchers if 'EngineSet' in l.__name__]
return [ProfileDir] + eslaunchers
n = Int(
num_cpus(),
config=True,
help=
"""The number of engines to start. The default is to use one for each
CPU on your machine""")
engine_launcher_class = DottedObjectName(
'LocalEngineSetLauncher',
config=True,
help="""The class for launching a set of Engines. Change this value
to use various batch systems to launch your engines, such as PBS,SGE,MPIExec,etc.
Each launcher class has its own set of configuration options, for making sure
it will work in your environment.
You can also write your own launcher, and specify it's absolute import path,
as in 'mymodule.launcher.FTLEnginesLauncher`.
Examples include:
LocalEngineSetLauncher : start engines locally as subprocesses [default]
MPIExecEngineSetLauncher : use mpiexec to launch in an MPI environment
PBSEngineSetLauncher : use PBS (qsub) to submit engines to a batch queue
SGEEngineSetLauncher : use SGE (qsub) to submit engines to a batch queue
SSHEngineSetLauncher : use SSH to start the controller
Note that SSH does *not* move the connection files
around, so you will likely have to do this manually
unless the machines are on a shared file system.
WindowsHPCEngineSetLauncher : use Windows HPC
""")
daemonize = Bool(
False,
config=True,
help="""Daemonize the ipcluster program. This implies --log-to-file.
Not available on Windows.
""")
def _daemonize_changed(self, name, old, new):
if new:
self.log_to_file = True
aliases = Dict(engine_aliases)
flags = Dict(engine_flags)
_stopping = False
def initialize(self, argv=None):
super(IPClusterEngines, self).initialize(argv)
self.init_signal()
self.init_launchers()
def init_launchers(self):
self.engine_launcher = self.build_launcher(self.engine_launcher_class)
self.engine_launcher.on_stop(lambda r: self.loop.stop())
def init_signal(self):
# Setup signals
signal.signal(signal.SIGINT, self.sigint_handler)
def build_launcher(self, clsname):
"""import and instantiate a Launcher based on importstring"""
if '.' not in clsname:
# not a module, presume it's the raw name in apps.launcher
clsname = 'IPython.parallel.apps.launcher.' + clsname
# print repr(clsname)
try:
klass = import_item(clsname)
except (ImportError, KeyError):
self.log.fatal("Could not import launcher class: %r" % clsname)
self.exit(1)
launcher = klass(work_dir=u'.', config=self.config, log=self.log)
return launcher
def start_engines(self):
self.log.info("Starting %i engines" % self.n)
self.engine_launcher.start(self.n, self.profile_dir.location)
def stop_engines(self):
self.log.info("Stopping Engines...")
if self.engine_launcher.running:
d = self.engine_launcher.stop()
return d
else:
return None
def stop_launchers(self, r=None):
if not self._stopping:
self._stopping = True
self.log.error("IPython cluster: stopping")
self.stop_engines()
# Wait a few seconds to let things shut down.
dc = ioloop.DelayedCallback(self.loop.stop, 4000, self.loop)
dc.start()
def sigint_handler(self, signum, frame):
self.log.debug("SIGINT received, stopping launchers...")
self.stop_launchers()
def start_logging(self):
# Remove old log files of the controller and engine
if self.clean_logs:
log_dir = self.profile_dir.log_dir
for f in os.listdir(log_dir):
if re.match(r'ip(engine|controller)z-\d+\.(log|err|out)', f):
os.remove(os.path.join(log_dir, f))
# This will remove old log files for ipcluster itself
# super(IPBaseParallelApplication, self).start_logging()
def start(self):
"""Start the app for the engines subcommand."""
self.log.info("IPython cluster: started")
# First see if the cluster is already running
# Now log and daemonize
self.log.info('Starting engines with [daemon=%r]' % self.daemonize)
# TODO: Get daemonize working on Windows or as a Windows Server.
if self.daemonize:
if os.name == 'posix':
daemonize()
dc = ioloop.DelayedCallback(self.start_engines, 0, self.loop)
dc.start()
# Now write the new pid file AFTER our new forked pid is active.
# self.write_pid_file()
try:
self.loop.start()
except KeyboardInterrupt:
pass
except zmq.ZMQError as e:
if e.errno == errno.EINTR:
pass
else:
raise
