loghole.in 76.3 KB
Newer Older
1 2 3 4
#! /usr/bin/env python

#
# EMULAB-COPYRIGHT
5
# Copyright (c) 2004, 2005, 2006, 2007 University of Utah and the Flux Group.
6 7 8 9 10 11
# All rights reserved.
#

import re
import pwd
import sys
12
import glob
13 14 15 16 17 18
import sets
import stat
import time
import getopt
import socket
import zipfile
19
import tempfile
20
import cStringIO
21
import threading
22 23 24 25 26 27 28 29 30 31 32
import xmlrpclib
import os, os.path
import ConfigParser

prefix = "@prefix@"

TBPATH = os.path.join(prefix, "lib")
if TBPATH not in sys.path:
    sys.path.append(TBPATH)
    pass

33 34
from tbevent import *

35 36 37 38 39 40 41
PACKAGE_VERSION = 0.1

# Default server
XMLRPC_SERVER = "@BOSSNODE@"
XMLRPC_PORT   = 3069

RSYNC = "@RSYNC@"
42
ACCEPTABLE_RSYNC_STATUS = [ 0 ]
43
NONRETRIABLE_RSYNC_STATUS = [ 0, 1, 2, 4 ]
44 45 46 47 48 49 50 51

DIRS = {
    "proj" : "proj",
    "groups" : "groups",
    }

EXPDIR_FMT = os.path.join("/", DIRS["proj"], "%(PID)s", "exp", "%(EID)s")

52 53
GLOBAL_LOGS = [
    "event-sched.log", "feedback.log", "../tbdata/feedback_data.tcl",
54
    "digest.out", "report.mail", "../tbdata/%(EID)s.report"
55
    ]
56

57 58
EXTENSIONS = [ ".zip" ]

59
LINK_DUMP_FMT = [ "trace_%s-%s.xmit", "trace_%s-%s.recv" ]
60

61 62
CLEANFILE = ".cleanonsync"

63 64 65
DOT_LOGHOLES = ""

DEFAULT_LOGHOLES = ["/var/emulab/logs", "/local/logs"]
66

67 68
DELAYLOG2XPLOT = os.path.join(prefix, "bin", "delaylog2xplot")

69 70
MERGECAP = "/usr/X11R6/bin/mergecap"
TCPTRACE = "/usr/local/bin/tcptrace"
71
TEVC     = os.path.join(prefix, "bin", "tevc")
72

73 74 75
PID = None
EID = None

76 77 78
COLLAB_PW = None
DPDBNAME  = None

79 80 81 82
try:
    pw = pwd.getpwuid(os.getuid())
    pass
except KeyError:
83 84 85 86
    sys.stderr.write("error: unknown user id %d" % os.getuid())
    sys.exit(2)
    pass

87 88
USER = pw.pw_name
HOME = pw.pw_dir
89 90 91

CERTIFICATE = os.path.join(HOME, ".ssl", "emulab.pem")

92
PARALLELIZATION = 25
93

94 95 96 97 98 99 100 101 102
DEBUG = 0

VERBOSITY_MACHINE = 1
VERBOSITY_HUMAN = 2

VERBOSITY = VERBOSITY_HUMAN

SERVER = None

103 104 105 106 107
# For compressfile stuff
COMPRESSFILE_THRESHOLD  = 512 * 1024
COMPRESSFILE_EXTENSIONS = [ ".zip", ".gz", ".bz2" ]
COMPRESSFILE_PROGRAM    = "/usr/bin/gzip"

108 109
OSIDS = {}

110 111 112 113
JOB_TRIES = 3
JOB_PROGRESS = [ "-", "\\", "|", "/" ]
JOB_PROGRESS_INDEX = 0

114 115 116 117 118 119 120 121 122 123
if len(XMLRPC_SERVER) == 0:
    sys.stderr.write("internal error: XMLRPC_SERVER is empty\n")
    sys.exit(2)
    pass

if len(RSYNC) == 0:
    sys.stderr.write("internal error: RSYNC is empty\n")
    sys.exit(2)
    pass

124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142
def loghole_include(path):
    retval = ""

    prefix = "/"
    for comp in path.split("/"):
        if comp != "":
            prefix = prefix + comp + "/"
            retval = retval + "+ " + prefix + "\n"
            pass
        pass

    retval = retval + "+ " + prefix + "**\n"
    
    return retval

for path in DEFAULT_LOGHOLES:
    DOT_LOGHOLES += loghole_include(path)
    pass

143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
##
# Initialize the XML-RPC server proxy, if it has not already been done.  After
# returning, the SERVER variable will hold the proxy object.
#
def init_server():
    global SERVER

    if not SERVER:
        if not os.path.exists(CERTIFICATE):
            sys.stderr.write("error: missing emulab certificate: %s\n" %
                             CERTIFICATE)
            sys.exit(2)
            pass
        
        from M2Crypto.m2xmlrpclib import SSL_Transport
        from M2Crypto import SSL
        
        URI = "https://" + XMLRPC_SERVER + ":" + str(XMLRPC_PORT) + prefix
        
        ctx = SSL.Context("sslv23")
        ctx.load_cert(CERTIFICATE, CERTIFICATE)
        ctx.set_verify(SSL.verify_none, 16)
        ctx.set_allow_unknown_ca(0)
        
        SERVER = xmlrpclib.ServerProxy(URI, SSL_Transport(ctx))
        pass
    return

##
# Download the user's experiment list from the XML-RPC server.
#
# @return A list of (pid, eid) pairs for each experiment created by the user.
#
def get_explist():
    global SERVER

    from emulabclient import RESPONSE_SUCCESS
    
    retval = []
    
    init_server()
    
    list_method = getattr(SERVER, "experiment.getlist")
    elist = list_method(PACKAGE_VERSION, {})
    
    if elist["code"] != RESPONSE_SUCCESS:
        sys.stderr.write("error: " + elist["output"] + "\n")
        sys.exit(1)
        pass
    else:
        for project in elist["value"]:
            for group in elist["value"][project]:
                for exp in elist["value"][project][group]:
                    retval.append((project, exp))
                    pass
                pass
            pass
        pass
    
    return retval

##
# Print the top level usage statement for the whole tool.
#
def usage():
208
    print "Usage: loghole [-hV] [-e pid/eid|-a] [OPTIONS] <action> [...]"
209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230
    print
    print "Manage an experiment's log hole."
    print
    print "Optional arguments:"
    print "  -h, --help                Print this message, or the action's"
    print "                            usage message."
    print "  -V, --version             Print version information."
    print "  -d, --debug               Turn on debugging messages."
    print "  -q, --quiet               Be less verbose."
    print "  -v, --verbose             Be more verbose."
    print "  -e, --experiment=PID/EID  Specify the experiment to manage."
    print "  -a, --all                 Perform the action on all of your "
    print "                            experiments."
    print "  -s, --server=SERVER       Specify the XML-RPC server to use."
    print "                            (Default: %s)" % XMLRPC_SERVER
    print "  -P, --port=PORT           Specify the XML-RPC server port."
    print "                            (Default: %d)" % XMLRPC_PORT
    print
    print "Actions:"
    print "  sync      Synchronize the experiment's log holes with the"
    print "            experiment's log directory."
    print "  archive   Archive the contents of the experiment's log directory."
231
    print "  change    Change the metadata of an existing archive."
232 233 234 235
    print "  list      Print a brief listing of the archives in the"
    print "            experiment's log directory."
    print "  show      Show a detailed listing of the archive(s) in the"
    print "            experiment directory."
236
    print "  validate  Validate the experiment logs."
237 238
    print "  clean     Clean the experiment log directory and node local"
    print "            directories."
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
    print "  gc        Garbage collect old archives."
    print
    print "Environment:"
    print "  cwd       The project and experiment ID will be inferred from the"
    print "            current working directory, if it is inside the"
    print "            experiment's directory (e.g. /proj/foo/exp/bar)."
    print
    print "Examples:"
    print "  $ loghole -e foo/bar sync"
    print "  $ loghole -e foo/bar list"
    return

##
# Print the usage statement for the "sync" action.
#
def sync_usage():
255
    print "Usage: loghole sync [-nPsR] [-r remotedir] [-l localdir] [node1 ...]"
256
    print
257
    print "Synchronize the experiment's log holes with a local directory."
258 259
    print
    print "Optional arguments:"
260 261 262 263 264 265 266 267 268 269 270
    print "  -r, --remote=REMOTEDIR"
    print "               An additional remote directory to sync.  This option"
    print "               is additive, so you can download several additional"
    print "               directories."
    print "  -l, --local=LOCALDIR"
    print "               The local directory to store the downloaded files."
    print "               This defaults to the experiment's log directory."
    print "  -n, --no-standard"
    print "               Flag that indicates that the standard logholes"
    print "               (i.e. '/var/emulab/logs', '/local/logs') should not"
    print "               be downloaded."
271 272
    print "  -P, --no-post"
    print "               Do not do any additional post processing."
273 274
    print "  -s, --snapshot"
    print "               Send a SNAPSHOT event to any link trace agents."
275 276
    print "  -R, --root"
    print "               Use sudo on remote nodes to read files as root."
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
    print "  node1 ...    The names of the virtual or physical nodes that "
    print "               should be synced.  The default behavior is to sync "
    print "               all of the nodes in an experiment."
    print
    print "Files:"
    print "  /proj/<pid>/exp/<eid>/logs/<node>/<loghole>"
    print "               The location of the log hole for a particular node."
    print
    print "Examples:"
    print "  $ loghole sync"
    return

##
# Print the usage statement for the "archive" action.
#
def archive_usage():
    print ("Usage: loghole archive [-k (i-delete|space-is-needed)] "
           "[-a days] [-c comment] [-d] [archive-name]")
    print
    print "Archive the contents of the experiment's log directory.  The"
    print "archive will be a zip file containing all of the directories"
    print "under the experiment's log directory, any extra log files"
    print "at the top (e.g. event-sched.log), and some metadata."
    print
    print "Optional arguments:"
    print "  -k, --keep-until=(i-delete|space-is-needed)"
    print "                         Keep the archive until you decide to"
    print "                         delete it manually or space is needed."
    print "                         (Default: space-is-needed)"
    print "  -a, --keep-atleast=N   Keep the archive atleast N days after"
    print "                         creation.  This value keeps the archive "
    print "                         from being garbage collected when more "
    print "                         space is needed. (Default: 3 days)"
    print "  -c, --comment=COMMENT  Add a comment to the archive. This comment"
    print "                         will be displayed in the listing generated"
    print "                         by the 'show' action.  Use a single dash "
    print "                         (-) to read the comment from standard in."
314 315
    print "  -d, --delete           Mark the experiment as clean-on-sync to"
    print "                         remove stale data before another sync."
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385
    print
    print "Required arguments:"
    print "  archive-name  The name of the archive to create."
    print "                (Default: The experiment name)"
    print
    print "Files:"
    print "  /proj/<pid>/exp/<eid>/logs/<archive-name>.<instance>.zip"
    print
    print "Examples:"
    print "  $ loghole archive foobar"
    return

##
# Print the usage statement for the "list" action.
#
def list_usage():
    print "Usage: loghole list [-O1!Xo] [-m atmost] [-s megabytes]"
    print
    print "Print a brief listing of the archives in the experiment's log"
    print "directory.  The listing can be helpful in getting an idea of"
    print "what archives are in your experiment and when they are likely"
    print "to be garbage collected."
    print
    print "Optional arguments:"
    print "  -O                    Only list archives that are marked as 'keep"
    print "                        until \"i-delete\"'."
    print "  -1                    Only list archives that are a day from"
    print "                        their keep-atleast date."
    print "  -!                    Only list archives that are past their"
    print "                        keep-atleast date."
    print "  -X                    Only list archives that are ready to be"
    print "                        garbage collected."
    print "  -o                    List archives that do not match the above"
    print "                        flags."
    print "  -m, --keep-atmost=N   Keep at most N archives in the experiment."
    print "                        (Default: 100)"
    print "  -s, --keep-size=N     Keep at N megabytes of archives."
    print "                        (Default: 3.0 MB)"
    print
    print "Examples:"
    print "  $ loghole list"
    print "  GC  Name             Date"
    print
    print "  [ ] foobar.1.zip     10/15"
    print "  [!] foobar.0.zip     10/13"
    print "  ..."
    print
    print "  $ loghole -q list -!"
    print "  foobar.0.zip"
    return

##
# Print the usage statement for the "show" action.
#
def show_usage():
    print "Usage: loghole show [archive-name]"
    print
    print "Show a detailed listing of the archive(s) in the experiment"
    print "directory.  The listing shows all of the metadata for the archive,"
    print "comments, and the file listing."
    print
    print "Example:"
    print "  $ loghole show foobar"
    print "  Archive: foobar.0.zip"
    print "    Creator:\t\t%s" % USER
    print "    ..."
    print "  Archive: foobar.1.zip"
    print "    ..."
    return

386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
##
# Print the usage statement for the "change" action.
#
def change_usage():
    print ("Usage: loghole change [-k (i-delete|space-is-needed)] "
           "[-a days] [-c comment] <archive-name1> [archive-name2 ...]")
    print
    print "Change the metadata of an existing archive."
    print
    print "Optional arguments:"
    print "  -k, --keep-until=(i-delete|space-is-needed)"
    print "                         Keep the archive until you decide to"
    print "                         delete it manually or space is needed."
    print "                         (Default: space-is-needed)"
    print "  -a, --keep-atleast=N   Keep the archive atleast N days after"
    print "                         creation.  This value keeps the archive "
    print "                         from being garbage collected when more "
    print "                         space is needed. (Default: 3 days)"
    print "  -c, --comment=COMMENT  Add a comment to the archive. This comment"
    print "                         will be displayed in the listing generated"
    print "                         by the 'show' action.  Use a single dash "
    print "                         (-) to read the comment from standard in."
    print
    print "Required arguments:"
    print "  archive-name  The name of the archive to change."
    print
    print "Files:"
    print "  /proj/<pid>/exp/<eid>/logs/<archive-name>.<instance>.zip"
    print
    print "Examples:"
    print "  $ loghole change -c 'Results bad...' foobar"
    return

##
# Print the usage statement for the "validate" action.
#
def validate_usage():
423
    print "Usage: loghole validate"
424
    print
425
    print "Validate the experiment logs.  Currently, this action will:"
426
    print
427 428 429
    print "  1. Examine the log files from individual logs from the "
    print "     program-agent to verify that the files are intact."
    print "  2. Check that the referent of any soft-links exist."
430 431
    print
    print "Examples:"
432
    print "  $ loghole validate"
433 434
    return

435 436 437 438
##
# Print the usage statement for the "clean" action.
#
def clean_usage():
439
    print "Usage: loghole clean [-fneR] [node1 ...]"
440 441 442 443
    print
    print "Clean the experiment log directory by removing any subdirectories"
    print "and/or clean the log holes on the nodes.  The default behavior is"
    print "to clean both."
444
    print
445 446 447 448
    print "Optional arguments:"
    print "  -f, --force         Do not ask for confirmation."
    print "  -n, --nodes         Clean only the nodes log directories."
    print "  -e, --experiment    Clean only the experiment log directory."
449
    print "  -R, --root          Use sudo to clean nodes as root."
450 451 452 453 454
    print
    print "Examples:"
    print "  $ loghole clean"
    return

455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470
##
# Print the usage statement for the "gc" action.
#
def gc_usage():
    print "Usage: loghole gc [-n] [-m atmost] [-s megabytes]"
    print
    print "Garbage collect old log archives."
    print
    print "Optional arguments:"
    print "  -n, --dry-run         Perform a dry run, do not delete any files."
    print "  -m, --keep-atmost=N   Keep at most N archives in the experiment."
    print "                        (Default: 100)"
    print "  -s, --keep-size=N     Keep at N megabytes of archives."
    print "                        (Default: 3.0 MB)"
    print
    print "Examples:"
471
    print "  $ loghole gc -n -m 1"
472 473 474
    print "  Reclaimed 3421 bytes from foo/bar."
    print "  52434 bytes are still in foo/bar."
    print
475
    print "  $ loghole -q gc -m 1"
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519
    print "  3421 52434"
    return

##
# Infer the project and experiment IDs from the given list of directories.
#
# @param *args The list of directories to check for likely PID/EID values.
# @return A pair containing the project and experiment IDs, which will be None
# if the IDs could not be inferred.
#
def infer_pid_eid(*args):
    pid = None
    eid = None
    for path in args:
        dirs = path.split(os.path.sep)
        if ((len(dirs) < 6) or not
            ((DIRS["proj"] in dirs and
              "exp" in dirs[dirs.index(DIRS["proj"]):]) or
             (len(dirs) >= 6 and
              (DIRS["groups"] in dirs and
               "exp" in dirs[dirs.index(DIRS["groups"]):])))):
            # No match...
            pass
        else:
            if DIRS["proj"] in dirs:
                pid = dirs[dirs.index(DIRS["proj"]) + 1]
                pass
            else:
                pid = dirs[dirs.index(DIRS["groups"]) + 1]
                pass
            eid = dirs[dirs.index("exp") + 1]
            pass
        pass

    return pid, eid

##
# Attempt to open a loghole archive located at the given path.  An archive
# must be a zip file with a "loghole.cfg" file stored in the root directory.
#
# @param path The path to the loghole archive.
# @return A pair containing the ZipFile object referring to the archive and a
# ConfigParser object that contains the archive's metadata.
#
520
def open_loghole(path, mode='r'):
521 522
    retval = None, None
    if zipfile.is_zipfile(path):
523
        zf = zipfile.ZipFile(path, mode)
524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772
        if "loghole.cfg" in zf.namelist():
            cio = cStringIO.StringIO(zf.read("loghole.cfg"))
            cp = ConfigParser.ConfigParser()
            cp.readfp(cio)
            retval = zf, cp
            zf = None
            pass
        if zf:
            zf.close()
            pass
        pass
    
    return retval

##
# Generate a name that is unique in the context of the given list of entries.
#
# @param base The base of the file name.
# @param ext The file name extension.
# @param entries An iterable containing any existing file names.
# @return A string of the form "<base>.<unique #>.<ext>".
#
def uniquish_name(base, ext, entries):
    retval = None
    lpc = 0
    while not retval:
        name = base + "." + str(lpc) + "." + ext
        if name not in entries:
            retval = name
            pass
        lpc = lpc + 1
        pass
    
    return retval

##
# Callback for os.walk() that writes files to the ZipFile.
#
# @param zf The ZipFile object that files should be added to.
# @param dirname The current directory name.
# @param names The list of files in the current directory.
#
def file_archiver(zf, dirname, names):
    for entry in names:
        entry = os.path.join(dirname, entry)
        if os.path.isfile(entry):
            zf.write(entry)
            pass
        pass

    return

##
# Sort the list of log archives in the current directory based on when they
# should be garbage collected.
#
# @param keep_atmost The maximum number of archives that should be kept, any
# extras are marked as collectable.
# @param keep_size The number of megabytes of archives to keep.
# @return A dictionary containing: "del", a list of archives that will be
# deleted at the next GC; "safe", a list of archives that are marked "keep
# until 'i-delete'"; "soon", the remaining archives; "reclaimed-size", the
# amount of data that will be reclaimed at the next GC; and "inuse-size", the
# amount of data that will remain after the next GC.
#
def sort_archives(keep_atmost, keep_size):
    
    keep_size = keep_size * 1024 * 1024

    reclaimed_size = 0
    inuse_size = 0 # In-use size for non-gc'able archives
    inuse_gc_size = 0 # In-use size for gc'able archives
    
    current_time = time.time()
    
    retval_del = []
    retval_safe = []
    retval_soon = []
    exp_logholes = []

    # Walk the directory collecting archives and sorting them into "safe" and
    # to-be-determined lists.
    for entry in os.listdir(os.curdir):
        st = os.stat(entry)
        lh, cp = open_loghole(entry)
        if lh:
            keep_until = cp.get("MAIN", "keep-until")
            if keep_until == "i-delete":
                if DEBUG:
                    print "loghole: ignoring - " + entry
                    pass

                retval_safe.append(entry)
                inuse_size = inuse_size + st[stat.ST_SIZE]
                pass
            else:
                created = time.mktime(time.strptime(cp.get("MAIN", "created")))
                keep_atleast = (cp.getint("MAIN", "keep-atleast") *
                                60 * 60 * 24)

                inuse_gc_size = inuse_gc_size + st[stat.ST_SIZE]
                exp_logholes.append((entry, st, created + keep_atleast))
                pass
            lh.close()
            pass
        pass

    # Sort the archives from oldest keep-atleast time to newest, then
    exp_logholes.sort(lambda x, y: cmp(x[2], y[2]))

    # ... divide them between the del list and the soon list depending on
    # whether or not space is needed.
    for entry, st, expiration in exp_logholes:
        if inuse_gc_size > keep_size:
            if DEBUG:
                print "loghole: stale entry - " + entry
                pass

            retval_del.append(entry)
            inuse_gc_size = inuse_gc_size - st[stat.ST_SIZE]
            reclaimed_size = reclaimed_size + st[stat.ST_SIZE]
            pass
        else:
            retval_soon.append(entry)
            pass
        pass

    # Resort the list in the reverse direction for keep-atleast time.
    exp_logholes = [x for x in exp_logholes if x[0] in retval_soon]
    exp_logholes.sort(lambda x, y: cmp(y[2], x[2]))

    # Update the inuse_size for the archives below the keep-atmost value, then
    for entry, st, expiration in exp_logholes[0:keep_atmost]:
        if DEBUG:
            print ("loghole: keeping '"
                   + entry
                   + "' until "
                   + time.ctime(expiration))
            pass

        inuse_size = inuse_size + st[stat.ST_SIZE]
        pass

    # ... move the rest to the "del" list and
    for entry, st, expiration in exp_logholes[keep_atmost:]:
        if DEBUG:
            print "loghole: excess entry - " + entry
            pass

        retval_soon.remove(entry)
        retval_del.append(entry)
        reclaimed_size = reclaimed_size + st[stat.ST_SIZE]
        pass

    # ... drop the end of the array.
    del retval_soon[keep_atmost:]

    return {
        "del" : retval_del,
        "safe" : retval_safe,
        "soon" : retval_soon,
        "reclaimed-size" : reclaimed_size,
        "inuse-size" : inuse_size,
        }

##
# Performs the "archive" action, which involves building a zip file containing
# all of the logs.
#
# @param args Action-specific command line arguments.
#
def do_archive(args):
    retval = 0

    keep_until = "space-is-needed"
    keep_atleast = 3
    comments = []
    delete_dir = False

    # Parse action-specific options and arguments.
    try:
        opts, args = getopt.getopt(args, "k:a:c:d", [
            "keep-until=", "keep-atleast=", "comment=", "delete" ])
        for opt, val in opts:
            if opt in ("-k", "--keep-until"):
                if val == "i-delete" or val == "space-is-needed":
                    keep_until = val
                    pass
                else:
                    raise getopt.error(
                        'error: expecting "i-delete" or "space-is-needed" for '
                        'keep until value, got - ' + val)
                pass
            elif opt in ("-a", "--keep-atleast"):
                try:
                    keep_atleast = int(val)
                    pass
                except:
                    raise getopt.error(
                        'error: expecting integer for keep atleast value, '
                        'got - ' + val)
                pass
            elif opt in ("-c", "--comment"):
                if val == "-":
                    comments.append(sys.stdin.read())
                    pass
                else:
                    comments.append(val)
                    pass
                pass
            elif opt in ("-d", "--delete"):
                delete_dir = True
                pass
            pass
        
        if len(args) > 1:
            raise getopt.error(
                'error: too many arguments for "archive" action.')
        pass
    except getopt.error, e:
        print e.args[0]
        archive_usage()
        return 2

    if len(args) < 1:
        archive_base = EID
        pass
    else:
        archive_base = args[0]
        pass

    logdir = os.path.join(EXPDIR_FMT % { "PID" : PID, "EID" : EID }, "logs")
    os.chdir(logdir)

    # Figure out a unique name for the archive,
    archive_name = uniquish_name(archive_base, 'zip', os.listdir(logdir))

    # ... create a new zip file,
    zf = zipfile.ZipFile(archive_name, 'w', zipfile.ZIP_DEFLATED)

    # ... fill it up with the node log holes, and
    for entry in os.listdir(logdir):
        if os.path.isdir(entry):
            os.path.walk(entry, file_archiver, zf)
            pass
        pass

    # ... any global log files (e.g. event-sched.log).
    for log in GLOBAL_LOGS:
773 774 775
        formatted_log = log % { "PID" : PID, "EID" : EID }
        if os.path.exists(formatted_log):
            zf.write(formatted_log)
776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822
            pass
        pass

    # Construct the meta-data.
    cp = ConfigParser.ConfigParser()
    cp.add_section("MAIN")
    cp.set("MAIN", "created", time.ctime(time.time()))
    cp.set("MAIN", "creator", USER)
    cp.set("MAIN", "project", PID)
    cp.set("MAIN", "experiment", EID)
    cp.set("MAIN", "keep-until", keep_until)
    cp.set("MAIN", "keep-atleast", keep_atleast)
    
    cio = cStringIO.StringIO()
    cp.write(cio)

    loghole_zi = zipfile.ZipInfo()
    loghole_zi.filename = "loghole.cfg"
    ct = time.localtime(time.time())
    loghole_zi.date_time = (ct[0], ct[1], ct[2], ct[3], ct[4], ct[5])
    loghole_zi.compress_type = zipfile.ZIP_STORED
    loghole_zi.comment = "Log hole meta-data"
    
    zf.writestr(loghole_zi, cio.getvalue())

    # Add any comments.
    ct = time.localtime(time.time())
    lpc = 0
    for comment in comments:
        loghole_zi = zipfile.ZipInfo()
        loghole_zi.filename = "loghole-comment.%d.txt" % lpc
        loghole_zi.date_time = (ct[0], ct[1], ct[2], ct[3], ct[4], ct[5])
        loghole_zi.compress_type = zipfile.ZIP_DEFLATED
        loghole_zi.comment = "Loghole comment #%d" % lpc

        zf.writestr(loghole_zi, comment)

        lpc = lpc + 1
        pass
    
    if DEBUG:
        zf.printdir()
        pass
    
    zf.close()

    if delete_dir:
823
        open(CLEANFILE, 'w')
824
        os.chmod(CLEANFILE, 0664)
825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
        pass

    if VERBOSITY == VERBOSITY_MACHINE:
        print archive_name
        pass
    elif VERBOSITY >= VERBOSITY_HUMAN:
        print "Created log hole archive: " + archive_name
        pass
    
    return retval

##
# Performs the "show" action, which shows a detailed listing of a log archive.
#
# @param args Action-specific command line arguments.
#
def do_show(args):
    retval = 0

    logdir = os.path.join(EXPDIR_FMT % { "PID" : PID, "EID" : EID }, "logs")
    os.chdir(logdir)
    
    for entry in os.listdir(logdir):
        if len(args) == 0 or entry.startswith(args[0]):
            lh, cp = open_loghole(entry)
            if lh:
                print "Archive: " + entry
                print ("  Project/Experiment:\t%s/%s" %
                       (cp.get("MAIN", "project"),
                        cp.get("MAIN", "experiment")))
                print "  Creator:\t\t%s" % cp.get("MAIN", "creator")
                print "  Created:\t\t%s" % cp.get("MAIN", "created")
                print "  Keep-until:\t\t%s" % cp.get("MAIN", "keep-until")
                print ("  Keep-atleast:\t\t%s days" %
                       cp.get("MAIN", "keep-atleast"))
860 861 862 863
                if "report.mail" in lh.namelist():
                    print "Report:"
                    print lh.read("report.mail")
                    pass
864 865 866 867 868 869 870 871 872
                for name in lh.namelist():
                    if re.match(r'loghole-comment.*\.txt', name):
                        before = 38 - (len(name) / 2)
                        after = 77 - before - len(name)
                        print "%s %s %s" % ("-" * before, name, "-" * after)
                        print lh.read(name)
                        pass
                    pass
                print "=" * 79
873 874 875 876 877
                try:
                    lh.printdir()
                    pass
                except IOError:
                    pass
878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893
                lh.close()
                print
                pass
            pass
        pass
    
    return retval

##
# Run rsync for a given host and list of files.
#
# @param host The host to sync against.
# @param base The local directory where the downloaded files should be placed.
# @param *args The list of directories to sync.
# @retval The exit status of the rsync command-line.
#
894 895
def rsync(host, base, dlpath = None, paths = ["/"], port = None,
          output = False, delay = 0, as_root = False, relative = True):
896 897 898
    global RSYNC, HOME, DEBUG
    
    retval = 0
899
    for path in paths:
900 901 902 903 904
        if relative == True:
            fullpath = os.path.join("", base, path[1:], "")
        else:
            fullpath = base
            pass
905 906 907 908 909
        try:
            os.makedirs(fullpath)
            pass
        except OSError:
            pass
910
        cmd = [RSYNC,
911 912
               "-rtgoDl",
               "-z", # rsync compression, ssh compression turned off below
913
               "-v", # XXX
914 915 916 917
               "--copy-unsafe-links"]
        if dlpath != None:
            cmd.append("--include-from=" + dlpath)
            pass
918
        if VERBOSITY > VERBOSITY_HUMAN:
919
            cmd.append("-v")
920
            pass
921
        ssh_args = ("ssh -o BatchMode=yes -o StrictHostKeyChecking=no "
922
                    + "-o Compression=no "
923 924 925 926 927
                    + "-o UserKnownHostsFile="
                    + os.path.join(HOME, ".ssh", "emulab_known_hosts"))
        if port:
            ssh_args += " -p " + str(port)
            pass
928 929 930
        if as_root:
            cmd.append("--rsync-path=sudo rsync")
            pass
931
        cmd.extend([("--rsh=" + ssh_args),
932 933
                    host + ":" + path,
                    fullpath])
934
        if DEBUG:
935 936
            print `cmd`
            retval = -1
937 938
            pass
        else:
939 940 941 942 943
            retval = os.fork()
            if retval != 0:
                # Parent
                pass
            else:
944
                try:
945 946 947
                    if delay > 0:
                        time.sleep(delay)
                        pass
948 949 950 951 952 953 954 955 956
                    os.close(0)
                    os.open("/dev/null", os.O_RDONLY)
                    if output:
                        os.close(1)
                        fd = os.open(os.path.join(base, ".rsync.out"),
                                     os.O_CREAT|os.O_TRUNC|os.O_WRONLY,
                                     0664)
                        os.dup2(fd, 2)
                        pass
957 958 959 960 961 962 963 964
                    try:
                        os.execve(RSYNC, cmd, os.environ)
                    except OSError, e:
                        sys.stderr.write(
                            "error: unable to exec '%s' - %s\n"
                            "error: args=%s\n"
                            % (RSYNC, e.strerror, `cmd`))
                        pass
965 966 967 968 969
                    os._exit(127)
                    pass
                except:
                    sys.exit(1)
                    pass
970
                pass
971 972 973 974 975
            pass
        pass
    
    return retval

976
def rmstar(host, port, delay, as_root, *args):
977 978 979 980 981
    global HOME, DEBUG

    retval = 0
    for path in args:
        cmd = ["/usr/bin/ssh",
982
               "-o", "BatchMode=yes",
983 984
               "-o", "StrictHostKeyChecking=no",
               "-o", ("UserKnownHostsFile=" +
985 986 987 988
                      os.path.join(HOME, ".ssh", "emulab_known_hosts"))]
        if port:
            cmd.extend([ "-p", str(port) ])
            pass
989 990 991 992
        cmd.extend([ host ])
        if as_root:
            cmd.extend(["sudo"])
        cmd.extend(["rm", "-rf", path ])
993 994 995 996 997 998 999 1000 1001 1002
        if DEBUG:
            print `cmd`
            retval = -1
            pass
        else:
            retval = os.fork()
            if retval != 0:
                # Parent
                pass
            else:
1003 1004 1005
                if delay > 0:
                    time.sleep(delay)
                    pass
1006 1007 1008 1009 1010 1011 1012
                os.execve("/usr/bin/ssh", cmd, os.environ)
                os._exit(127)
                pass
            pass
        pass
    
    return retval
1013

1014 1015 1016 1017 1018 1019 1020 1021 1022
def print_status(msg):
    if not sys.stdout.isatty() or VERBOSITY < VERBOSITY_HUMAN:
        return
    
    sys.stdout.write(msg)
    sys.stdout.flush()
    
    return

1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
#
# Compress a file. Return False if file is skipped for whatever reason.
#
def compressfile(root, file, threshold):
    pathname = os.path.join(root, file)
    if not os.path.exists(pathname):
        return False
    statinfo = os.stat(pathname)
    mode     = statinfo[stat.ST_MODE]
    size     = statinfo[stat.ST_SIZE]
    # Only regular files
    if not stat.S_ISREG(mode):
        return False
    # Only files greater then threshold
    if not size >= threshold:
        return False
    first, last = os.path.splitext(pathname)
    # Try to avoid compressing things that are already compressed.
    if last in COMPRESSFILE_EXTENSIONS:
        return False
    # Lets be careful of foo and foo.gz, which is from a previous run.
    if os.path.exists(pathname + ".gz"):
        os.unlink(pathname + ".gz")
        pass
    print_status("Compressing " + pathname + " --> " + pathname + ".gz\n")
    exitval = os.system(COMPRESSFILE_PROGRAM + " %s " % (pathname,))
    if exitval != 0:
        return False
    return True

1053
def print_job_status(mode, hosts, jobs, hoststatus, hoststart):
1054 1055
    if not sys.stdout.isatty() or VERBOSITY < VERBOSITY_HUMAN:
        return
1056
    
1057
    lmsg = "nodes todo/doing/done %d/%d/%d "
1058
    rmsg = mode + " time min/avg/max: %.2f/%.2f/%.2f secs"
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078

    hosttimes = map(lambda x:
                    hoststatus.get(x, (time.time(),))[0] - hoststart[x],
                    hoststart.keys())
    
    if len(hosttimes) == 0:
        hosttimes = [0]
        avg = 0.0
        pass
    else:
        avg = sum(hosttimes) / len(hosttimes)
        pass
    lmsg = lmsg % (
        len(hosts) - len(hoststatus) - len(jobs),
        len(jobs),
        len(hoststatus))
    rmsg = rmsg % (
        min(hosttimes),
        avg,
        max(hosttimes))
1079
    line_len = len(lmsg) + len(rmsg) + 2
1080 1081 1082
    if line_len < 79:
        lmsg += " " * (79 - line_len)
        pass
1083 1084
    print_status(lmsg + rmsg + " "
                 + JOB_PROGRESS[JOB_PROGRESS_INDEX % len(JOB_PROGRESS)] + "\r")
1085 1086

    return
1087

1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
##
# Detect changes in the rsync log file sizes.
#
# @param jobs The hashtable of jobs
# @param logsize The hashtable of previous log file sizes.
# @param logdir The root directory for log files.
# @return True if the size of any of the log files has changed.
#
def making_progress(jobs, logsize, logdir):
    retval = False
    for pid in jobs:
        phost, vhost = jobs[pid]
        try:
            st = os.stat(os.path.join(logdir, vhost, ".rsync.out"))
            if st[stat.ST_SIZE] > logsize[pid]:
                retval = True
                logsize[pid] = st[stat.ST_SIZE]
                pass
            pass
        except OSError:
            pass
        pass
    
    return retval

1113 1114 1115 1116 1117 1118
##
# Trace preprocessing.
#
# @param traces Sequence of dictionaries describing the active link tracers.
#
def trace_pre(traces):
1119 1120
    if len(traces) == 0:
        return;
1121
    
1122 1123 1124 1125
    rc = os.system(TEVC + " -w -t 30 -e %s/%s now __all_tracemon snapshot" %
                   (PID, EID))
    if rc != 0:
        print "Could not SNAPSHOT the trace agents!";
1126 1127 1128 1129 1130 1131 1132 1133 1134
        pass
    return

##
# Trace post processing
#
# @param link2trace Dictionary containing the link names that map to a sequence
# of dictionaries containing individual trace information.
#
1135 1136
def trace_post(link2trace):
    lpc = 1
Leigh B. Stoller's avatar
Leigh B. Stoller committed
1137 1138
    sql = []
    
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
    for linkvname in link2trace.keys():
        print_status("linktrace post processing %d/%d  " %
                     (lpc, len(link2trace)))
        args = {}
        os.chdir(linkvname)
        for filename in glob.glob("*.recv*"):
            base, ext = os.path.splitext(filename)
            if base.startswith("__") or ext == ".recv":
                continue
            al = args.get(ext, [])
            al.append(filename)
            args[ext] = al
1151 1152 1153 1154 1155 1156

            stuff = base.split(".");
            trace_db = link2trace[linkvname][stuff[0]]["trace_db"]
            if trace_db != "0":
                sql.append(linkvname + "/" + filename)
                pass
1157 1158 1159 1160 1161 1162
            pass
        for ext in args.keys():
            if ext == ".recv":
                continue
            dst = "__all.recv%s" % (ext,)
            print_status(".")
1163 1164 1165
            if os.path.exists(MERGECAP):
                os.system(MERGECAP + " -w %s %s" % (dst, " ".join(args[ext])))
                pass
1166
            print_status(".")
1167 1168 1169 1170 1171 1172 1173
            if ext != ".0":
                if not os.path.exists(ext[1:]):
                    os.mkdir(ext[1:])
                    pass
                os.chdir(ext[1:])
                dst = os.path.join(os.path.pardir, dst)
                pass
1174 1175 1176
            if os.path.exists(TCPTRACE):
                os.system(TCPTRACE + " -y -G %s > .tcptrace.out" % (dst,))
                pass
1177 1178 1179
            if ext != ".0":
                os.chdir(os.path.pardir)
                pass
1180
            pass
1181
        print_status("\r")
1182 1183 1184 1185 1186 1187 1188
        os.chdir(os.path.pardir)
        lpc += 1
        pass

    if len(link2trace.keys()):
        print_status("\n")
        pass
1189

1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
    # Now do the DB enters.
    if len(sql) != 0:
        SNORTCONF  = ".snort.conf"
        SNORTOUT   = ".snort.out"
        USERRULES  = "snort.rules"
        
        print_status("Entering trace data into mysql DB %s" % DPDBNAME)

        # snort needs a config file to do anything useful
        os.system("/bin/rm -f %s %s" % (SNORTCONF, SNORTOUT))
        
        fp = open(SNORTCONF, 'w')
        os.chmod(SNORTCONF, 0600)
1203 1204
        fp.write("output database: log, mysql, user=%s dbname=%s password=%s\n" %
                 (USER, DPDBNAME, COLLAB_PW))
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216

        # Allow user to override default rules.
        if os.path.exists(USERRULES):
            fp.write("include ./%s" % (USERRULES,))
            pass
        else:
            fp.write("log ip any any -> any any\n")
            pass
        fp.close()
        
        for filename in sql:
            print_status(".")
1217
            os.system("snort -k none -l ./ -c ./%s -r %s > %s 2>&1" %
1218 1219 1220 1221
                      (SNORTCONF, filename, SNORTOUT))
            print_status(".")
            pass
        print_status("\n")
1222
        os.system("/bin/rm -f %s" % (SNORTCONF,))
1223 1224
        pass
    
1225 1226
    return

1227 1228 1229 1230
##
# Performs the "sync" action, which synchronizes the logholes for a group of
# nodes.
#
1231 1232
# XXX This method is in dire need of a rewrite.
#
1233 1234
# @param args Action-specific command line arguments.
#
1235
def do_sync(args, mode="sync", must_be_active=1, as_root=False):
1236
    from emulabclient import RESPONSE_SUCCESS
1237
    global OSIDS, SERVER, DOT_LOGHOLES, JOB_PROGRESS_INDEX
1238
    global DPDBNAME, COLLAB_PW
1239 1240
    
    retval = 0
1241
    delete_dir = False
1242

1243
    logdir = os.path.join(EXPDIR_FMT % { "PID" : PID, "EID" : EID }, "logs")
1244 1245
    # Special dir for local program agent (or other stuff).
    opsdir = os.path.join(EXPDIR_FMT % { "PID" : PID, "EID" : EID }, "ops")
1246 1247

    standard_logholes = True
1248
    snapshot_traces = False
1249
    post_process = True
1250
    compress = False
1251 1252 1253
    extra_logholes = []

    try:
1254 1255
        opts, node_list = getopt.getopt(args, "nl:r:PsdcR", [
            "no-standard", "local=", "remote=", "delete", "compress", "root"])
1256 1257 1258 1259
        for opt, val in opts:
            if opt in ("-n", "--no-standard"):
                standard_logholes = False
                pass
1260 1261 1262
            elif opt in ("-P", "--no-post"):
                post_process = False
                pass
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
            elif opt in ("-l", "--local"):
                logdir = os.path.abspath(val)
                if not os.path.isdir(logdir):
                    raise getopt.error(
                        'error: local directory does not exist - '
                        + logdir)
                pass
            elif opt in ("-r", "--remote"):
                extra_logholes.append(val)
                pass
1273 1274 1275
            elif opt in ("-s", "--snapshot"):
                snapshot_traces = True
                pass
1276 1277 1278
            elif opt in ("-d", "--delete"):
                delete_dir = True
                pass
1279 1280
            elif opt in ("-c", "--compress"):
                compress = True
1281 1282
            elif opt in ("-R", "--root"):
                as_root = True
1283
                pass
1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
            pass
        pass
    except getopt.error, e:
        print e.args[0]
        sync_usage()
        return 2

    if not standard_logholes:
        DOT_LOGHOLES = ""
        pass
    
    for path in extra_logholes:
        DOT_LOGHOLES += loghole_include(path)
        pass

    DOT_LOGHOLES += "- *\n"
    
1301 1302 1303
    init_server()
    
    state_method = getattr(SERVER, "experiment.state")
1304
    
1305
    os.chdir(logdir)
1306 1307
    
    if standard_logholes and os.path.exists(CLEANFILE):
1308 1309 1310 1311
        do_clean(["-f", "-e"])
        os.unlink(CLEANFILE)
        pass

1312 1313 1314
    dlhandle = tempfile.NamedTemporaryFile(prefix=".logholes")
    dlhandle.write(DOT_LOGHOLES)
    dlhandle.flush()
1315 1316 1317 1318 1319 1320

    state = state_method(PACKAGE_VERSION, { "proj" : PID, "exp" : EID })
    if state["code"] != RESPONSE_SUCCESS:
        sys.stderr.write("error: " + state["output"] + "\n")
        retval = 1
        pass
1321 1322 1323
    elif (state["value"] != "active" and
          state["value"] != "activating" and
          state["value"] != "modify_reswap"):
1324 1325 1326 1327 1328 1329 1330 1331
        if must_be_active:
            sys.stderr.write("error: %s/%s is not an active experiment!\n" %
                             (PID, EID))
            retval = 2
            pass
        else:
            retval = 0
            pass
1332 1333 1334
        pass
    else:
        info_method = getattr(SERVER, "experiment.info")
1335
        osinfo_method = getattr(SERVER, "osid.info")
1336
        ssh_method = getattr(SERVER, "node.sshdescription")
1337 1338
        collabpw_method = getattr(SERVER, "user.collabpassword")
        dpinfo_method = getattr(SERVER, "experiment.dpinfo")
1339

1340
        # Get any bits of information we need from the RPC server.
1341 1342 1343 1344 1345
        physical = info_method(PACKAGE_VERSION, {
            "proj" : PID,
            "exp" : EID,
            "aspect" : "mapping",
            })
1346

1347
        traces = info_method(PACKAGE_VERSION, {
1348 1349
            "proj" : PID,
            "exp" : EID,
1350
            "aspect" : "traces",
1351
            })
1352 1353 1354 1355 1356

        ssh_data = ssh_method(PACKAGE_VERSION, {
            "proj" : PID,
            "exp" : EID
            })
1357 1358 1359 1360

        collab_data = collabpw_method(PACKAGE_VERSION, { })
        dpinfo_data = dpinfo_method(PACKAGE_VERSION, { "proj" : PID,
                                                       "exp" : EID, })
1361
        
1362
        if ((physical["code"] != RESPONSE_SUCCESS) or
1363
            (traces["code"] != RESPONSE_SUCCESS) or
1364 1365
            (collab_data["code"] != RESPONSE_SUCCESS) or
            (dpinfo_data["code"] != RESPONSE_SUCCESS) or
1366
            (ssh_data["code"] != RESPONSE_SUCCESS)):
1367 1368 1369 1370 1371 1372
            sys.stderr.write("error: xmlrpc failed - %s\n" %
                             (physical["output"],))
            retval = 1
            pass
        else:
            physical = physical["value"]
1373
            traces = traces["value"]
1374
            ssh_data = ssh_data["value"]
1375 1376
            COLLAB_PW = collab_data["value"]
            DPDBNAME  = dpinfo_data["value"]["dbname"]
1377 1378 1379 1380

            if snapshot_traces:
                trace_pre(traces)
                pass
1381 1382

            # Separate the nodes into those that are up
1383
            hosts = [(x["node"], x["name"]) for x in physical.values()
1384 1385 1386
                     if ((x["erole"] == "virthost" or
                          x["erole"] == "node" or
                          x["erole"] == "delaynode") and
1387 1388
                         (x["status"] == "up" and x["eventstatus"] == "ISUP"))]

1389
            # ... or down.
1390 1391 1392 1393 1394 1395
            downhosts = [(x["node"], x["name"]) for x in physical.values()
                     if ((x["erole"] == "virthost" or
                          x["erole"] == "node" or
                          x["erole"] == "delaynode") and
                         (x["status"] != "up" or x["eventstatus"] != "ISUP"))]

1396 1397
            phosts = sets.Set([x[0] for x in hosts])
            vhosts = sets.Set([x[1] for x in hosts])
1398 1399
            dphosts = sets.Set([x[0] for x in downhosts])
            dvhosts = sets.Set([x[1] for x in downhosts])
1400 1401
            if len(node_list) == 0:
                node_list = vhosts
1402 1403 1404 1405 1406
                
                for host in downhosts:
                    sys.stderr.write("warning: '%s' is down, its logs will "
                                     "not be retrieved\n" % (host[1],))
                    pass
1407 1408
                pass

1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
            #
            # Do the local ops directory first.
            #
            if os.path.exists(opsdir):
                if mode == "sync":
                    pid = rsync("localhost", logdir,
                                paths = [opsdir], output = True,
                                relative = False)
                    pass
                elif mode == "clean":
                    pid = rmstar("localhost", None, 0, False, opsdir + "/*")
                    pass
                else:
                    assert 0
                pass

1425 1426
            known_nodes = sets.Set(node_list) & (phosts | vhosts)
            unknown_nodes = known_nodes ^ sets.Set(node_list)
1427

1428 1429
            jobs = {}
            hoststatus = {}
1430
            hoststart = {}
1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
            hosttries = {}
            logsize = {}
            while len(hosts) > 0 or len(jobs) > 0:
                if len(hosts) > 0:
                    ht = hosts.pop()
                    phost, vhost = ht
                    pass
                else:
                    phost = ""
                    vhost = ""
                    pass
1442
                if (vhost in node_list) or (phost in node_list):
1443
                    # Start a job for the next host.
1444 1445 1446
                    osid = physical[vhost]["osid"]
                    if osid not in OSIDS:
                        oi = osinfo_method(PACKAGE_VERSION, {
1447
                            "osid" : osid })
1448 1449 1450 1451 1452
                        if oi["code"] == RESPONSE_SUCCESS:
                            OSIDS[osid] = oi["value"]
                            pass
                        else:
                            OSIDS[osid] = None
1453 1454
                            pass
                        pass
1455

1456
                    # Make sure we can talk to it.
1457 1458 1459 1460 1461 1462
                    if not OSIDS[osid]:
                        sys.stderr.write("warning: unknown OSID %s for %s\n" %
                                         (physical[vhost]["osid"], vhost))
                        continue
                    else:
                        vos = OSIDS[physical[vhost]["osid"]]
1463 1464 1465
                        if ("osfeatures" not in vos or
                            not vos["osfeatures"] or
                            "ssh" not in vos["osfeatures"]):
1466 1467 1468 1469 1470 1471 1472 1473
                            if VERBOSITY > VERBOSITY_HUMAN:
                                sys.stderr.write("warning: %s OS, %s, does "
                                                 "not support ssh, skipping\n"
                                                 % (vhost, vos["osname"]))
                                pass
                            continue
                        pass

1474 1475 1476 1477 1478 1479 1480
                    # Start the job.
                    if ht in hosttries:
                        delay = 3
                        pass
                    else:
                        delay = 0
                        pass
1481
                    if mode == "sync":
1482
                        pid = rsync(phost, os.path.join(logdir, vhost),
1483
                                    dlpath = dlhandle.name,
1484
                                    port = ssh_data[phost].get("port", None),
1485
                                    output = True,
1486 1487
                                    delay = delay,
                                    as_root = as_root)
1488 1489
                        pass
                    elif mode == "clean":
1490 1491 1492
                        pid = rmstar(phost,
                                     ssh_data[phost].get("port", None),
                                     delay,
1493
                                     as_root,
1494
                                     "/local/logs/*")
1495 1496 1497 1498
                        pass
                    else:
                        assert 0
                    
1499 1500 1501
                    jobs[pid] = ht
                    logsize[pid] = 0
                    hoststart[ht] = time.time()
1502

1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
                    if ht not in hosttries:
                        hosttries[ht] = 0
                        pass
                    hosttries[ht] += 1
                    pass
                
                try:
                    # Wait for the children and update the status.
                    while len(jobs) == PARALLELIZATION or len(hosts) == 0:
                        print_job_status(mode, node_list, jobs,
                                         hoststatus, hoststart)
                        # Check for any changes in the log files.
                        if (mode == "sync" and
                            making_progress(jobs, logsize, logdir)):
                            JOB_PROGRESS_INDEX += 1
                            pass
                        # Check for processes that have ended.
                        pid, sts = os.waitpid(0, os.WNOHANG)
                        if pid in jobs:
                            # Retry on some errors 
                            if (mode == "clean" and sts == 255 and
                                hosttries[jobs[pid]] <= JOB_TRIES):
                                hosts.append(jobs[pid])
                                pass
                            elif (mode == "sync" and
                                  sts not in NONRETRIABLE_RSYNC_STATUS and
                                  hosttries[jobs[pid]] <= JOB_TRIES):
                                hosts.append(jobs[pid])
                                pass
                            else:
1533
                                hoststatus[jobs[pid]] = (time.time(), sts)
1534
                                pass
1535
                            del jobs[pid]
1536
                            pass
1537

1538 1539 1540
                        if len(jobs) == PARALLELIZATION or len(hosts) == 0:
                            time.sleep(0.5)
                            pass
1541
                        pass
1542
                    pass
1543 1544
                except OSError:
                    pass
1545 1546
                pass

1547 1548
            sys.stdout.write("\n")

1549 1550
            # Clue the user in if there were nodes that they specified nodes
            # we don't know about.
1551 1552
            if len(unknown_nodes) > 0:
                for vhost in unknown_nodes:
1553 1554 1555 1556 1557 1558 1559 1560
                    if vhost in dvhosts or vhost in dphosts:
                        sys.stderr.write("error: node '%s' is down\n" %
                                         (vhost,))
                        pass
                    else:
                        sys.stderr.write("error: unknown node - %s\n" %
                                         (vhost,))
                        pass
1561 1562 1563
                    pass
                retval = 2
                pass
1564

1565 1566 1567
            # Check the status of the jobs and dump logs for any that failed.
            for ht, (etime, sts) in hoststatus.items():
                phost, vhost = ht
1568
                if VERBOSITY > VERBOSITY_HUMAN:
1569
                    print "info: %s - %.2f" % (vhost, etime - hoststart[ht])
1570
                    pass
1571 1572
                if mode == "clean":
                    continue
1573
                if sts not in ACCEPTABLE_RSYNC_STATUS:
1574 1575
                    sys.stderr.write("error: failed to %s node '%s'\n" %
                                     (mode, vhost))
1576 1577 1578 1579 1580
                    eout = open(os.path.join(logdir, vhost, ".rsync.out"),
                                "r").read()
                    sys.stderr.write(
                        vhost + ": " + re.sub(r'\n',
                                              "\n" + vhost + ": ",
1581
                                              eout[:-1]) + eout[-1])
1582 1583
                    retval = 3
                    pass
1584
                else:
1585 1586 1587 1588 1589
                    # Post process the delay-agent log files, if there are any.
                    vdir = os.path.join(vhost, "var", "emulab", "logs")
                    if (os.path.exists(vdir) and
                        os.path.exists("delayagent.debug")):
                        os.chdir(vdir)
1590 1591 1592 1593 1594
                        os.system("%s %s" %
                                  (DELAYLOG2XPLOT, "delayagent.debug"))
                        pass
                    os.chdir(logdir)
                    pass
1595
                pass
1596 1597

            # Add symbolic links to the trace files.
1598
            link2trace = {}
1599
            for trace in traces:
1600 1601
                if trace["trace_type"] == "monitor":
                    continue
1602 1603 1604

                tl = link2trace.get(trace["linkvname"], {})
                tl[trace["vnode"]] = trace
1605 1606
                link2trace[trace["linkvname"]] = tl
                
1607 1608 1609 1610 1611 1612 1613 1614
                fullpath = os.path.join(logdir, trace["linkvname"])
                try:
                    os.makedirs(fullpath)
                    pass
                except OSError:
                    pass
                for fmt in LINK_DUMP_FMT:
                    src = os.path.join(trace["delayvname"],
1615 1616
                                       "local",
                                       "logs",
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633
                                       fmt % (trace["vnode"],
                                              trace["linkvname"]))
                    for filename in glob.glob(src + "*"):
                        dst = os.path.split(filename)
                        dst = dst[len(dst) - 1].split('.')
                        comp = [trace["vnode"]]
                        comp.extend(dst[1:])
                        dst = ".".join(comp)
                        dst = os.path.join(fullpath, dst)
                        try:
                            os.unlink(dst)
                            pass
                        except OSError:
                            pass
                        os.symlink(os.path.join(os.pardir, filename), dst)
                        pass
                    pass
1634

1635 1636
                # Add some links to the xplot log files created by
                # delaylog2xplot
1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651
                src = os.path.join(trace["delayvname"],
                                   "var", "emulab", "logs",
                                   "%s-%s-*.xpl" % (trace["linkvname"],
                                                    trace["vnode"]))
                for filename in glob.glob(src):
                    aspect = filename.split("-")[2].split(".")[0]
                    dst = "%s-%s.xpl" % (trace["vnode"], aspect)
                    dst = os.path.join(fullpath, dst)
                    try:
                        os.unlink(dst)
                        pass
                    except OSError:
                        pass
                    os.symlink(os.path.join(os.pardir, filename), dst)
                    pass