def main(): namespace = parse_job_mode() if namespace.jmx_username: namespace.jmx_password = getpass.getpass(prompt="JMX Password ") else: namespace.jmx_password = None if bool(namespace.seed_host) + bool(namespace.host) + bool( namespace.host_file) != 1: error( "Exactly one of --seed-host, --host and --host-file must be used", print_traceback=False) hosts = None if namespace.host_file: with open(namespace.host_file) as f: hosts = f.readlines() if namespace.host: hosts = namespace.host cstar.output.configure(namespace.verbose) with cstar.job.Job() as job: env = {} job_id = str(uuid.uuid4()) msg("Job id is", emph(job_id)) cstar.signalhandler.print_message_and_save_on_sigint(job, job_id) job.setup(hosts=hosts, seeds=namespace.seed_host, command=namespace.command, job_id=job_id, strategy=cstar.strategy.parse( fallback(namespace.strategy, "topology")), cluster_parallel=fallback(namespace.cluster_parallel, False), dc_parallel=fallback(namespace.dc_parallel, False), max_concurrency=namespace.max_concurrency, timeout=namespace.timeout, env=env, stop_after=namespace.stop_after, job_runner=cstar.jobrunner.LocalJobRunner, key_space=namespace.key_space, output_directory=namespace.output_directory, ignore_down_nodes=False, dc_filter=namespace.dc_filter, sleep_on_new_runner=namespace.ssh_pause_time, sleep_after_done=namespace.node_done_pause_time, ssh_username=namespace.ssh_username, ssh_password=namespace.ssh_password, ssh_identity_file=namespace.ssh_identity_file, ssh_lib=namespace.ssh_lib, jmx_username=namespace.jmx_username, jmx_password=namespace.jmx_password, resolve_hostnames=namespace.resolve_hostnames, hosts_variables=namespace.hosts_variables) job.run()
def print_message_and_save_on_sigint(job, job_id): global _msg, _job _msg = "Shutting down gracefully. Hit ^C again to shut down gracelessly.\n\nTo resume, type %s" % ( emph("cstar continue " + job_id), ) _job = job signal.signal(signal.SIGINT, _handler)
def execute_command(args): cstar.output.debug(args) command = args.command if bool(args.seed_host) + bool(args.host) + bool(args.host_file) != 1: error( "Exactly one of --seed-host, --host and --host-file must be used", print_traceback=False) hosts = None if args.host_file: with open(args.host_file) as f: hosts = f.readlines() if args.host: hosts = args.host with cstar.job.Job() as job: env = dict( (arg.name, getattr(args, arg.name)) for arg in command.arguments) if bool(args.enforced_job_id) == 1: job_id = args.enforced_job_id if not (validate_uuid4(job_id)): raise BadArgument("Job id is not a valid UUID v4 value.") else: job_id = str(uuid.uuid4()) msg("Job id is", emph(job_id)) msg("Running", command.file) cstar.signalhandler.print_message_and_save_on_sigint(job, job_id) job.setup(hosts=hosts, seeds=args.seed_host, command=command.file, job_id=job_id, strategy=cstar.strategy.parse( fallback(args.strategy, command.strategy, "topology")), cluster_parallel=fallback(args.cluster_parallel, command.cluster_parallel, False), dc_parallel=fallback(args.dc_parallel, command.dc_parallel, False), max_concurrency=args.max_concurrency, timeout=args.timeout, env=env, stop_after=args.stop_after, job_runner=cstar.jobrunner.RemoteJobRunner, key_space=args.key_space, output_directory=args.output_directory, ignore_down_nodes=args.ignore_down_nodes, dc_filter=args.dc_filter, sleep_on_new_runner=args.ssh_pause_time, sleep_after_done=args.node_done_pause_time, ssh_username=args.ssh_username, ssh_password=args.ssh_password, ssh_identity_file=args.ssh_identity_file, ssh_lib=args.ssh_lib, jmx_username=args.jmx_username, jmx_password=args.jmx_password, jmx_passwordfile=args.jmx_passwordfile) job.run()
def print_outcome(self): if self.state.is_done() and not self.errors: if len(self.state.progress.done) == self.state.stop_after: cstar.jobwriter.write(self) msg("Job", self.job_id, "successfully ran on", self.state.stop_after, "hosts.\nTo finish the job, run", emph("cstar continue %s" % (self.job_id,))) msg("Job", self.job_id, "finished successfully") else: msg("Job", self.job_id, "finished with errors.\n" "%s nodes finished successfully\n" "%s nodes had errors\n" "%s nodes didn't start executing" % (len(self.state.progress.done), len(self.state.progress.failed), len(self.state.original_topology) - len(self.state.progress.done) - len(self.state.progress.failed)))