mpdboot.py

fortran并行计算包

#!/usr/bin/env python
#
#   (C) 2001 by Argonne National Laboratory.
#       See COPYRIGHT in top-level directory.
#

"""
usage:  mpdboot --totalnum=<n_to_start> [--file=<hostsfile>]  [--help] \ 
                [--rsh=<rshcmd>] [--user=<user>] [--mpd=<mpdcmd>]      \ 
                [--loccons] [--remcons] [--shell] [--verbose] [-1]     \
                [--ncpus=<ncpus>] [--ifhn=<ifhn>] [--chkup] [--chkuponly] \
                [--maxbranch=<maxbranch>]
 or, in short form, 
        mpdboot -n n_to_start [-f <hostsfile>] [-h] [-r <rshcmd>] [-u <user>] \ 
                [-m <mpdcmd>]  -s -v [-1] [-c]

--totalnum specifies the total number of mpds to start; at least
  one mpd will be started locally, and others on the machines specified
  by the file argument; by default, only one mpd per host will be
  started even if the hostname occurs multiple times in the hosts file
-1 means remove the restriction of starting only one mpd per machine; 
  in this case, at most the first mpd on a host will have a console
--file specifies the file of machines to start the rest of the mpds on;
  it defaults to mpd.hosts
--mpd specifies the full path name of mpd on the remote hosts if it is
  not in your path
--rsh specifies the name of the command used to start remote mpds; it
  defaults to ssh; an alternative is rsh
--shell says that the Bourne shell is your default for rsh' 
--verbose shows the ssh attempts as they occur; it does not provide
  confirmation that the sshs were successful
--loccons says you do not want a console available on local mpd(s)
--remcons says you do not want consoles available on remote mpd(s)
--ncpus indicates how many cpus you want to show for the local machine;
  others are listed in the hosts file
--ifhn indicates the interface hostname to use for the local mpd; others
  may be specified in the hostsfile
--chkup requests that mpdboot try to verify that the hosts in the host file
  are up before attempting start mpds on any of them; it just checks the number
  of hosts specified by -n
--chkuponly requests that mpdboot try to verify that the hosts in the host file
  are up; it then terminates; it just checks the number of hosts specified by -n
--maxbranch indicates the maximum number of mpds to enter the ring under another;
  the default is 4
"""
from time import ctime
__author__ = "Ralph Butler and Rusty Lusk"
__date__ = ctime()
__version__ = "$Revision: 1.49 $"
__credits__ = ""

import re

from os       import environ, path, kill, access, X_OK
from sys      import argv, exit, stdout
from popen2   import Popen4, Popen3, popen2
from socket   import gethostname, gethostbyname_ex
from select   import select, error
from signal   import SIGKILL
from commands import getoutput, getstatusoutput
from mpdlib   import mpd_set_my_id, mpd_get_my_username, mpd_same_ips, \
                     mpd_get_ranks_in_binary_tree, mpd_print, MPDSock

global myHost, fullDirName, rshCmd, user, mpdCmd, debug, verbose

def mpdboot():
    global myHost, fullDirName, rshCmd, user, mpdCmd, debug, verbose
    myHost = gethostname()
    mpd_set_my_id('mpdboot_%s' % (myHost) )
    fullDirName  = path.abspath(path.split(argv[0])[0])
    rshCmd = 'ssh'
    user = mpd_get_my_username()
    mpdCmd = path.join(fullDirName,'mpd.py')
    hostsFilename = 'mpd.hosts'
    totalnumToStart = 1    # may get chgd below
    debug = 0
    verbose = 0
    localConArg  = ''
    remoteConArg = ''
    oneMPDPerHost = 1
    myNcpus = 1
    myIfhn = ''
    chkupIndicator = 0  # 1 -> chk and start ; 2 -> just chk
    maxUnderOneRoot = 4
    try:
        shell = path.split(environ['SHELL'])[-1]
    except:
        shell = 'csh'

    argidx = 1    # skip arg 0
    while argidx < len(argv):
        if   argv[argidx] == '-h' or argv[argidx] == '--help':
            usage()
        elif argv[argidx] == '-r':    # or --rsh=
            rshCmd = argv[argidx+1]
            argidx += 2
        elif argv[argidx].startswith('--rsh'):
            splitArg = argv[argidx].split('=')
            try:
                rshCmd = splitArg[1]
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx] == '-u':    # or --user=
            user = argv[argidx+1]
            argidx += 2
        elif argv[argidx].startswith('--user'):
            splitArg = argv[argidx].split('=')
            try:
                user = splitArg[1]
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx] == '-m':    # or --mpd=
            mpdCmd = argv[argidx+1]
            argidx += 2
        elif argv[argidx].startswith('--mpd'):
            splitArg = argv[argidx].split('=')
            try:
                mpdCmd = splitArg[1]
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx] == '-f':    # or --file=
            hostsFilename = argv[argidx+1]
            argidx += 2
        elif argv[argidx].startswith('--file'):
            splitArg = argv[argidx].split('=')
            try:
                hostsFilename = splitArg[1]
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx].startswith('--ncpus'):
            splitArg = argv[argidx].split('=')
            try:
                myNcpus = int(splitArg[1])
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx].startswith('--ifhn'):
            splitArg = argv[argidx].split('=')
            myIfhn = splitArg[1]
            myHost = splitArg[1]
            argidx += 1
        elif argv[argidx] == '-n':    # or --totalnum=
            totalnumToStart = int(argv[argidx+1])
            argidx += 2
        elif argv[argidx].startswith('--totalnum'):
            splitArg = argv[argidx].split('=')
            try:
                totalnumToStart = int(splitArg[1])
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx].startswith('--maxbranch'):
            splitArg = argv[argidx].split('=')
            try:
                maxUnderOneRoot = int(splitArg[1])
            except:
                print 'mpdboot: invalid argument:', argv[argidx]
                usage()
            argidx += 1
        elif argv[argidx] == '-d' or argv[argidx] == '--debug':
            debug = 1
            argidx += 1
        elif argv[argidx] == '-s' or argv[argidx] == '--shell':
            shell = 'bourne'
            argidx += 1
        elif argv[argidx] == '-v' or argv[argidx] == '--verbose':
            verbose = 1
            argidx += 1
        elif argv[argidx] == '-c' or argv[argidx] == '--chkup':
            chkupIndicator = 1
            argidx += 1
        elif argv[argidx] == '--chkuponly':
            chkupIndicator = 2
            argidx += 1
        elif argv[argidx] == '-1':
            oneMPDPerHost = 0
            argidx += 1
        elif argv[argidx] == '--loccons':
            localConArg  = '-n'
            argidx += 1
        elif argv[argidx] == '--remcons':
            remoteConArg = '-n'
            argidx += 1
        else:
            print 'mpdboot: unrecognized argument:', argv[argidx]
            usage()
    if debug:
        print 'debug: starting'

    lines = []
    if totalnumToStart > 1:
        try:
            f = open(hostsFilename,'r')
            for line in f:
                lines.append(line)
        except:
            print 'unable to open (or read) hostsfile %s' % (hostsFilename)
            exit(-1)
    hostsAndInfo = [ {'host' : myHost, 'ncpus' : myNcpus, 'ifhn' : myIfhn} ]
    for line in lines:
        line = line.strip()
        if not line  or  line[0] == '#':
            continue
        splitLine = re.split(r'\s+',line)
        host = splitLine[0]
        ncpus = 1  # default
        if ':' in host:
            (host,ncpus) = host.split(':',1)
            ncpus = int(ncpus)
        ifhn = ''  # default
        for kv in splitLine[1:]:
            (k,v) = kv.split('=',1)
            if k == 'ifhn':
                ifhn = v
        hostsAndInfo.append( {'host' : host, 'ncpus' : ncpus, 'ifhn' : ifhn} )
    cachedIPs = {}
    if oneMPDPerHost  and  totalnumToStart > 1:
        oldHostsAndInfo = hostsAndInfo[:]
        hostsAndInfo = []
        for hostAndInfo in oldHostsAndInfo:
            oldhost = hostAndInfo['host']
            try: