def main() :
    #----------------------- Defining utility command-line interface -------------------------
    import timestamps_upload_options
    import data_upload_options
    import study_book_options
    import data_seeding_options

    an_usage_description = "%prog"
    an_usage_description += timestamps_upload_options.usage_description()
    an_usage_description += study_book_options.usage_description()
    an_usage_description += data_upload_options.usage_description()
    
    from cloudflu import VERSION
    a_version = "%s" % VERSION
    
    from optparse import IndentedHelpFormatter
    a_help_formatter = IndentedHelpFormatter( width = 127 )
    
    from optparse import OptionParser
    an_option_parser = OptionParser( usage = an_usage_description, version = a_version, formatter = a_help_formatter )
    

    #----------------------- Definition of the command line arguments ------------------------
    timestamps_upload_options.add( an_option_parser )

    study_book_options.add( an_option_parser )

    data_upload_options.add( an_option_parser )

    data_seeding_options.add( an_option_parser )

    common.concurrency_options.add( an_option_parser )

    common.communication_options.add( an_option_parser )

    amazon.security_options.add( an_option_parser )

    common.options.add( an_option_parser )

    
    #------------------ Extracting and verifying command-line arguments ----------------------
    an_options, an_args = an_option_parser.parse_args()

    common.options.extract( an_option_parser )
    
    AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY = amazon.security_options.extract( an_option_parser )
    
    common.communication_options.extract( an_option_parser )

    a_number_threads = common.concurrency_options.extract( an_option_parser )
    
    an_upload_seed_size = data_seeding_options.extract( an_option_parser )

    a_study_name = study_book_options.extract( an_option_parser )
    
    a_booked = data_upload_options.extract( an_option_parser )

    a_solver_run, a_case_dir, a_solver_log, a_watched_keyword, a_time_log = timestamps_upload_options.extract( an_option_parser )
    

    print_d( "\n--------------------------- Canonical substitution ------------------------\n" )
    import sys; an_engine = sys.argv[ 0 ]

    a_call = "%s %s %s %s %s" % ( an_engine, 
                                  timestamps_upload_options.compose( a_solver_run, a_case_dir, a_solver_log, a_watched_keyword, a_time_log ),
                                  data_upload_options.compose( a_booked ),
                                  study_book_options.compose( a_study_name ),
                                  data_seeding_options.compose( an_upload_seed_size ) )
    print_d( a_call + '\n' )


    print_d( "\n----------------------- Running actual functionality ----------------------\n" )
    import pexpect

    a_spent_time = common.Timer()

    import os.path; a_watched_dir = os.path.join( a_case_dir, 'processor0' )
    import os; a_listdir = os.listdir( a_watched_dir )
    
    from cloudflu.amazon.s3 import TRootObject
    a_root_object = TRootObject.get( AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY )
    print_d( "a_root_object = %s\n" % a_root_object )

    from cloudflu.common import WorkerPool
    a_worker_pool = WorkerPool( a_number_threads )
    
    a_child = pexpect.spawn( a_solver_run )
    
    import os.path; a_log_file = os.path.join( a_case_dir, a_solver_log )
    a_child.logfile = file( a_log_file, 'w' )

    while True :
        try:
            try:
                a_child.expect( a_watched_keyword )
            except:
                break
            
            import os; a_list_dir = os.listdir( a_watched_dir )

            for an_item in a_list_dir :
                if an_item in a_listdir :
                    continue
                
                a_worker_pool.charge( upload_timedata, ( a_root_object, a_study_name, a_booked,
                                                         an_upload_seed_size, a_number_threads, 
                                                         a_case_dir, an_item ) )
                if a_time_log == True :
                    a_worker_pool.charge( upload_timelog, ( a_root_object, a_study_name, a_booked,
                                                            an_upload_seed_size, a_number_threads, 
                                                            an_item, a_log_file ) )
                    pass
                pass
            
            a_listdir = a_list_dir
        except:
            import sys, traceback
            traceback.print_exc( file = sys.stderr )
            break
        
        pass

    a_worker_pool.shutdown()
    a_worker_pool.join()

    print_d( "a_spent_time = %s, sec\n" % a_spent_time )


    print_d( "\n------------------ Printing succussive pipeline arguments -----------------\n" )
    print a_study_name


    print_d( "\n--------------------------- Canonical substitution ------------------------\n" )
    print_d( a_call + '\n' )
    

    print_d( "\n-------------------------------------- OK ---------------------------------\n" )
    import sys, os; sys.exit( os.EX_OK )
    pass
Exemplo n.º 2
0
def main() :
    #----------------------- Defining utility command-line interface -------------------------
    import data_upload_options
    import study_book_options
    import upload_start_options
    import data_seeding_options

    an_usage_description = "%prog"
    an_usage_description += upload_start_options.usage_description()
    an_usage_description += data_upload_options.usage_description()
    
    from cloudflu import VERSION
    a_version = "%s" % VERSION
    
    from optparse import IndentedHelpFormatter
    a_help_formatter = IndentedHelpFormatter( width = 127 )
    
    from optparse import OptionParser
    an_option_parser = OptionParser( usage = an_usage_description, version = a_version, formatter = a_help_formatter )
    

    #----------------------- Definition of the command line arguments ------------------------
    data_upload_options.add( an_option_parser )

    upload_start_options.add( an_option_parser )

    study_book_options.add( an_option_parser )

    data_seeding_options.add( an_option_parser )

    common.concurrency_options.add( an_option_parser )

    common.communication_options.add( an_option_parser )

    amazon.security_options.add( an_option_parser )
    
    common.options.add( an_option_parser )

    
    #------------------ Extracting and verifying command-line arguments ----------------------
    an_options, an_args = an_option_parser.parse_args()

    common.options.extract( an_option_parser )
    
    AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY = amazon.security_options.extract( an_option_parser )
    
    common.communication_options.extract( an_option_parser )

    a_number_threads = common.concurrency_options.extract( an_option_parser )
    
    an_upload_seed_size = data_seeding_options.extract( an_option_parser )

    a_file2locations = upload_start_options.extract( an_option_parser )

    a_booked = data_upload_options.extract( an_option_parser )
    
    a_study_name = study_book_options.extract( an_option_parser )
    

    print_i( "--------------------------- Defining the study object ---------------------------\n" )
    from cloudflu.amazon.s3 import TRootObject
    a_root_object = TRootObject.get( AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY )
    print_d( "a_root_object = %s\n" % a_root_object )
    

    print_i( "-------------------------- Running actual functionality -------------------------\n" )
    a_study_object = entry_point( a_root_object, a_file2locations, an_upload_seed_size, a_number_threads, a_study_name, a_booked )
    
    
    print_i( "-------------------- Printing succussive pipeline arguments ---------------------\n" )
    print a_study_object.name()

    
    print_i( "-------------------------------------- OK ---------------------------------------\n" )
    pass
Exemplo n.º 3
0
def main() :
    #----------------------- Defining utility command-line interface -------------------------    
    import data_transfer_options
    import solver_start_options
    import data_seeding_options
    import openmpi_config_options

    an_usage_description = "%prog"
    an_usage_description += ec2.use.options.usage_description()
    an_usage_description += data_transfer_options.usage_description()
    an_usage_description += solver_start_options.usage_description()

    from cloudflu import VERSION
    a_version = "%s" % VERSION

    from optparse import IndentedHelpFormatter
    a_help_formatter = IndentedHelpFormatter( width = 127 )

    from optparse import OptionParser
    an_option_parser = OptionParser( usage = an_usage_description, version = a_version, formatter = a_help_formatter )


    #----------------------- Definition of the command line arguments ------------------------
    ec2.use.options.add( an_option_parser )
    
    data_transfer_options.add( an_option_parser )

    solver_start_options.add( an_option_parser )

    data_seeding_options.add( an_option_parser )

    common.concurrency_options.add( an_option_parser )

    openmpi_config_options.add( an_option_parser )

    amazon.security_options.add( an_option_parser )
    
    common.options.add( an_option_parser )
  
 
    #------------------ Extracting and verifying command-line arguments ----------------------
    an_options, an_args = an_option_parser.parse_args()
    
    an_enable_debug, a_log_file = common.options.extract( an_option_parser )
    
    AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY = amazon.security_options.extract( an_option_parser )
    
    a_cluster_id = ec2.use.options.extract( an_option_parser )

    a_hostfile = openmpi_config_options.extract( an_option_parser )

    a_number_threads = common.concurrency_options.extract( an_option_parser )

    an_upload_seed_size = data_seeding_options.extract( an_option_parser )

    a_case_dir, a_run_hook = solver_start_options.extract( an_option_parser )

    a_study_name = data_transfer_options.extract( an_option_parser )
    
    from cloudflu.preferences import get
    a_cluster_location = get( 'amazon.cluster.location' )
    a_host_port = int( get( 'amazon.cluster.host_port' ) )
    a_login_name = get( 'amazon.cluster.login_name' )


    print_d( "\n--------------------------- Canonical substitution ------------------------\n" )
    import sys; an_engine = sys.argv[ 0 ]

    a_call = "%s %s %s %s %s" % ( an_engine,
                                  ec2.use.options.compose( a_cluster_id ),
                                  data_transfer_options.compose( a_study_name ), 
                                  solver_start_options.compose( a_case_dir, a_run_hook ),
                                  openmpi_config_options.compose( a_hostfile ) )
    print_d( a_call + '\n' )


    print_d( "\n----------------------- Running actual functionality ----------------------\n" )
    a_spent_time = Timer()

    an_ec2_conn = ec2.common.region_connect( a_cluster_location, AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY )
    a_reservation = ec2.use.get_reservation( an_ec2_conn, a_cluster_id )
    print_d( '< %r > : %s\n' % ( a_reservation, a_reservation.instances ) )

    a_master_node = an_instance = a_reservation.instances[ 0 ]


    print_d( "\n--------------------- Uploading case data to S3 ---------------------------\n"  )
    import os; a_case_name = os.path.basename( a_case_dir )

    from cloudflu.amazon.s3 import TRootObject
    a_root_object = TRootObject.get( AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY )
    print_d( "a_root_object = %s\n" % a_root_object )
    
    import data_upload
    an_input_study = data_upload.entry_point( a_root_object, { a_case_dir : '' }, an_upload_seed_size, a_number_threads )

    an_input_study_name = an_input_study.name()
    print_d( "an_input_study_name = '%s'\n" % an_input_study_name )
    

    print_d( "\n------------------ Installing cloudflu to master node ----------------------\n"  )
    a_password = None
    a_host_name = an_instance.public_dns_name
    an_identity_file = ec2.run.get_identity_filepath( an_instance.key_name )
    
    ssh.options.echo( a_password, an_identity_file, a_host_port, a_login_name, a_host_name )
    a_ssh_client = ssh.connect( a_password, an_identity_file, a_host_port, a_login_name, a_host_name )
    
    an_instance2ssh = {}
    an_instance2ssh[ an_instance ] = a_ssh_client
    

    print_d( "\n------------- Downloading case data from S3 to the master node ------------\n"  )
    a_stdout_lines = ssh.command( a_ssh_client, 'python -c "import os, os.path, tempfile; print tempfile.mkdtemp()"' )
    a_working_dir = a_stdout_lines[ 0 ][ : -1 ]
    print_d( "a_working_dir = '%s'\n" % a_working_dir )

    ssh.command( a_ssh_client, "cloudflu-download --study-name=%s --output-dir=%s --remove" % ( an_input_study_name, a_working_dir ) )
    
    print_d( "\n--- Sharing the solver case folder for all the cluster nodes through NFS --\n" )
    ssh.command( a_ssh_client, "sudo sh -c 'echo %s *\(rw,no_root_squash,sync,subtree_check\) >> /etc/exports'" % ( a_working_dir ) )
    ssh.command( a_ssh_client, "sudo exportfs -a" ) # make changes effective on the running NFS server

    for an_instance in a_reservation.instances[ 1 : ] :
        a_host_name = an_instance.public_dns_name
        ssh.options.echo( a_password, an_identity_file, a_host_port, a_login_name, a_host_name )
        
        a_ssh_client = ssh.connect( a_password, an_identity_file, a_host_port, a_login_name, a_host_name )
        ssh.command( a_ssh_client, "mkdir -p %s" % ( a_working_dir ) )
        ssh.command( a_ssh_client, "sudo mount %s:%s %s" % ( a_master_node.private_ip_address, a_working_dir, a_working_dir ) )
        
        an_instance2ssh[ an_instance ] = a_ssh_client
        pass
    
    print_d( "\n------------------------ Get access the booked study ----------------------\n" )
    import study_book; 
    an_output_study = study_book.entry_point( a_root_object, a_study_name, True )


    print_d( "\n----------------------- Running of the solver case ------------------------\n" )
    a_num_nodes = len( a_reservation.instances )

    a_target_dir = os.path.join( a_working_dir, a_case_name )

    an_additional_args = "--study-name='%s' --booked --upload-seed-size=%d" % ( a_study_name, an_upload_seed_size )

    a_ssh_client = an_instance2ssh[ a_master_node ]
    ssh.command( a_ssh_client, "%s/%s %d '%s' %s" % ( a_target_dir, a_run_hook, a_num_nodes, a_hostfile, an_additional_args ) )

    an_output_study.seal()

    [ a_ssh_client.close() for a_ssh_client in an_instance2ssh.values() ]
    
    print_d( "a_spent_time = %s, sec\n" % a_spent_time )
    

    print_d( "\n------------------ Printing succussive pipeline arguments -----------------\n" )
    ec2.use.options.track( a_cluster_id )

    
    print_d( "\n--------------------------- Canonical substitution ------------------------\n" )
    print_d( a_call + '\n' )
    

    print_d( "\n-------------------------------------- OK ---------------------------------\n" )
    pass