Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Fetching contributors…

Octocat-spinner-32-eaf2f5

Cannot retrieve contributors at this time

file 617 lines (569 sloc) 25.174 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616
# Copyright (c) 2010-2011 OpenStack, LLC.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import os
from os.path import basename, dirname, isdir, join
import random
import shutil
import time
import logging
import hashlib
import itertools
import cPickle as pickle
import errno
import uuid

import eventlet
from eventlet import GreenPool, tpool, Timeout, sleep, hubs
from eventlet.green import subprocess
from eventlet.support.greenlets import GreenletExit

from swift.common.ring import Ring
from swift.common.utils import whataremyips, unlink_older_than, lock_path, \
        compute_eta, get_logger, write_pickle, renamer, dump_recon_cache, \
        TRUE_VALUES
from swift.common.bufferedhttp import http_connect
from swift.common.daemon import Daemon

hubs.use_hub('poll')

PICKLE_PROTOCOL = 2
ONE_WEEK = 604800
HASH_FILE = 'hashes.pkl'


def quarantine_renamer(device_path, corrupted_file_path):
    """
In the case that a file is corrupted, move it to a quarantined
area to allow replication to fix it.

:params device_path: The path to the device the corrupted file is on.
:params corrupted_file_path: The path to the file you want quarantined.

:returns: path (str) of directory the file was moved to
:raises OSError: re-raises non errno.EEXIST / errno.ENOTEMPTY
exceptions from rename
"""
    from_dir = dirname(corrupted_file_path)
    to_dir = join(device_path, 'quarantined', 'objects', basename(from_dir))
    invalidate_hash(dirname(from_dir))
    try:
        renamer(from_dir, to_dir)
    except OSError, e:
        if e.errno not in (errno.EEXIST, errno.ENOTEMPTY):
            raise
        to_dir = "%s-%s" % (to_dir, uuid.uuid4().hex)
        renamer(from_dir, to_dir)
    return to_dir


def hash_suffix(path, reclaim_age):
    """
Performs reclamation and returns an md5 of all (remaining) files.

:param reclaim_age: age in seconds at which to remove tombstones
"""
    md5 = hashlib.md5()
    for hsh in sorted(os.listdir(path)):
        hsh_path = join(path, hsh)
        try:
            files = os.listdir(hsh_path)
        except OSError, err:
            if err.errno == errno.ENOTDIR:
                partition_path = dirname(path)
                objects_path = dirname(partition_path)
                device_path = dirname(objects_path)
                quar_path = quarantine_renamer(device_path, hsh_path)
                logging.exception(
                    _('Quarantined %s to %s because it is not a directory') %
                    (hsh_path, quar_path))
                continue
            raise
        if len(files) == 1:
            if files[0].endswith('.ts'):
                # remove tombstones older than reclaim_age
                ts = files[0].rsplit('.', 1)[0]
                if (time.time() - float(ts)) > reclaim_age:
                    os.unlink(join(hsh_path, files[0]))
                    files.remove(files[0])
        elif files:
            files.sort(reverse=True)
            meta = data = tomb = None
            for filename in list(files):
                if not meta and filename.endswith('.meta'):
                    meta = filename
                if not data and filename.endswith('.data'):
                    data = filename
                if not tomb and filename.endswith('.ts'):
                    tomb = filename
                if (filename < tomb or # any file older than tomb
                    filename < data or # any file older than data
                    (filename.endswith('.meta') and
                     filename < meta)): # old meta
                    os.unlink(join(hsh_path, filename))
                    files.remove(filename)
        if not files:
            os.rmdir(hsh_path)
        for filename in files:
            md5.update(filename)
    try:
        os.rmdir(path)
    except OSError:
        pass
    return md5.hexdigest()


def invalidate_hash(suffix_dir):
    """
Invalidates the hash for a suffix_dir in the partition's hashes file.

:param suffix_dir: absolute path to suffix dir whose hash needs
invalidating
"""

    suffix = os.path.basename(suffix_dir)
    partition_dir = os.path.dirname(suffix_dir)
    hashes_file = join(partition_dir, HASH_FILE)
    with lock_path(partition_dir):
        try:
            with open(hashes_file, 'rb') as fp:
                hashes = pickle.load(fp)
            if suffix in hashes and not hashes[suffix]:
                return
        except Exception:
            return
        hashes[suffix] = None
        write_pickle(hashes, hashes_file, partition_dir, PICKLE_PROTOCOL)


def get_hashes(partition_dir, recalculate=[], do_listdir=False,
               reclaim_age=ONE_WEEK):
    """
Get a list of hashes for the suffix dir. do_listdir causes it to mistrust
the hash cache for suffix existence at the (unexpectedly high) cost of a
listdir. reclaim_age is just passed on to hash_suffix.

:param partition_dir: absolute path of partition to get hashes for
:param recalculate: list of suffixes which should be recalculated when got
:param do_listdir: force existence check for all hashes in the partition
:param reclaim_age: age at which to remove tombstones

:returns: tuple of (number of suffix dirs hashed, dictionary of hashes)
"""

    hashed = 0
    hashes_file = join(partition_dir, HASH_FILE)
    with lock_path(partition_dir):
        modified = False
        hashes = {}
        try:
            with open(hashes_file, 'rb') as fp:
                hashes = pickle.load(fp)
        except Exception:
            do_listdir = True
        if do_listdir:
            hashes = dict(((suff, hashes.get(suff, None))
                       for suff in os.listdir(partition_dir)
                       if len(suff) == 3 and isdir(join(partition_dir, suff))))
            modified = True
        for hash_ in recalculate:
            hashes[hash_] = None
        for suffix, hash_ in hashes.items():
            if not hash_:
                suffix_dir = join(partition_dir, suffix)
                if os.path.exists(suffix_dir):
                    try:
                        hashes[suffix] = hash_suffix(suffix_dir, reclaim_age)
                        hashed += 1
                    except OSError:
                        logging.exception(_('Error hashing suffix'))
                        hashes[suffix] = None
                else:
                    del hashes[suffix]
                modified = True
                sleep()
        if modified:
            write_pickle(hashes, hashes_file, partition_dir, PICKLE_PROTOCOL)
        return hashed, hashes


def tpooled_get_hashes(*args, **kwargs):
    """
Hack to work around Eventlet's tpool not catching and reraising Timeouts.
We return the Timeout, None if it's raised, the caller looks for it
and reraises it if found.
"""
    try:
        return get_hashes(*args, **kwargs)
    except Timeout, err:
        return err, None


class ObjectReplicator(Daemon):
    """
Replicate objects.

Encapsulates most logic and data needed by the object replication process.
Each call to .replicate() performs one replication pass. It's up to the
caller to do this in a loop.
"""

    def __init__(self, conf):
        """
:param conf: configuration object obtained from ConfigParser
:param logger: logging object
"""
        self.conf = conf
        self.logger = get_logger(conf, log_route='object-replicator')
        self.devices_dir = conf.get('devices', '/srv/node')
        self.mount_check = conf.get('mount_check', 'true').lower() in \
                              ('true', 't', '1', 'on', 'yes', 'y')
        self.vm_test_mode = conf.get(
                'vm_test_mode', 'no').lower() in ('yes', 'true', 'on', '1')
        self.swift_dir = conf.get('swift_dir', '/etc/swift')
        self.port = int(conf.get('bind_port', 6000))
        self.concurrency = int(conf.get('concurrency', 1))
        self.stats_interval = int(conf.get('stats_interval', '300'))
        self.object_ring = Ring(join(self.swift_dir, 'object.ring.gz'))
        self.ring_check_interval = int(conf.get('ring_check_interval', 15))
        self.next_check = time.time() + self.ring_check_interval
        self.reclaim_age = int(conf.get('reclaim_age', 86400 * 7))
        self.partition_times = []
        self.run_pause = int(conf.get('run_pause', 30))
        self.rsync_timeout = int(conf.get('rsync_timeout', 900))
        self.rsync_io_timeout = conf.get('rsync_io_timeout', '30')
        self.http_timeout = int(conf.get('http_timeout', 60))
        self.lockup_timeout = int(conf.get('lockup_timeout', 1800))
        self.recon_enable = conf.get(
                'recon_enable', 'no').lower() in TRUE_VALUES
        self.recon_cache_path = conf.get(
                'recon_cache_path', '/var/cache/swift')
        self.recon_object = os.path.join(self.recon_cache_path, "object.recon")

    def _rsync(self, args):
        """
Execute the rsync binary to replicate a partition.

:returns: return code of rsync process. 0 is successful
"""
        start_time = time.time()
        ret_val = None
        try:
            with Timeout(self.rsync_timeout):
                proc = subprocess.Popen(args, stdout=subprocess.PIPE,
                    stderr=subprocess.STDOUT)
                results = proc.stdout.read()
                ret_val = proc.wait()
        except Timeout:
            self.logger.error(_("Killing long-running rsync: %s"), str(args))
            proc.kill()
            return 1 # failure response code
        total_time = time.time() - start_time
        for result in results.split('\n'):
            if result == '':
                continue
            if result.startswith('cd+'):
                continue
            if not ret_val:
                self.logger.info(result)
            else:
                self.logger.error(result)
        if ret_val:
            self.logger.error(_('Bad rsync return code: %(args)s -> %(ret)d'),
                    {'args': str(args), 'ret': ret_val})
        elif results:
            self.logger.info(
                _("Successful rsync of %(src)s at %(dst)s (%(time).03f)"),
                {'src': args[-2], 'dst': args[-1], 'time': total_time})
        else:
            self.logger.debug(
                _("Successful rsync of %(src)s at %(dst)s (%(time).03f)"),
                {'src': args[-2], 'dst': args[-1], 'time': total_time})
        return ret_val

    def rsync(self, node, job, suffixes):
        """
Synchronize local suffix directories from a partition with a remote
node.

:param node: the "dev" entry for the remote node to sync with
:param job: information about the partition being synced
:param suffixes: a list of suffixes which need to be pushed

:returns: boolean indicating success or failure
"""
        if not os.path.exists(job['path']):
            return False
        args = [
            'rsync',
            '--recursive',
            '--whole-file',
            '--human-readable',
            '--xattrs',
            '--itemize-changes',
            '--ignore-existing',
            '--timeout=%s' % self.rsync_io_timeout,
            '--contimeout=%s' % self.rsync_io_timeout,
        ]
        if self.vm_test_mode:
            rsync_module = '[%s]::object%s' % (node['ip'], node['port'])
        else:
            rsync_module = '[%s]::object' % node['ip']
        had_any = False
        for suffix in suffixes:
            spath = join(job['path'], suffix)
            if os.path.exists(spath):
                args.append(spath)
                had_any = True
        if not had_any:
            return False
        args.append(join(rsync_module, node['device'],
                    'objects', job['partition']))
        return self._rsync(args) == 0

    def check_ring(self):
        """
Check to see if the ring has been updated

:returns: boolean indicating whether or not the ring has changed
"""
        if time.time() > self.next_check:
            self.next_check = time.time() + self.ring_check_interval
            if self.object_ring.has_changed():
                return False
        return True

    def update_deleted(self, job):
        """
High-level method that replicates a single partition that doesn't
belong on this node.

:param job: a dict containing info about the partition to be replicated
"""

        def tpool_get_suffixes(path):
            return [suff for suff in os.listdir(path)
                    if len(suff) == 3 and isdir(join(path, suff))]
        self.replication_count += 1
        begin = time.time()
        try:
            responses = []
            suffixes = tpool.execute(tpool_get_suffixes, job['path'])
            if suffixes:
                for node in job['nodes']:
                    success = self.rsync(node, job, suffixes)
                    if success:
                        with Timeout(self.http_timeout):
                            http_connect(node['ip'], node['port'],
                                node['device'], job['partition'], 'REPLICATE',
                                '/' + '-'.join(suffixes),
                          headers={'Content-Length': '0'}).getresponse().read()
                    responses.append(success)
            if not suffixes or (len(responses) == \
                        self.object_ring.replica_count and all(responses)):
                self.logger.info(_("Removing partition: %s"), job['path'])
                tpool.execute(shutil.rmtree, job['path'], ignore_errors=True)
        except (Exception, Timeout):
            self.logger.exception(_("Error syncing handoff partition"))
        finally:
            self.partition_times.append(time.time() - begin)

    def update(self, job):
        """
High-level method that replicates a single partition.

:param job: a dict containing info about the partition to be replicated
"""
        self.replication_count += 1
        begin = time.time()
        try:
            hashed, local_hash = tpool.execute(tpooled_get_hashes, job['path'],
                    do_listdir=(self.replication_count % 10) == 0,
                    reclaim_age=self.reclaim_age)
            # See tpooled_get_hashes "Hack".
            if isinstance(hashed, BaseException):
                raise hashed
            self.suffix_hash += hashed
            attempts_left = self.object_ring.replica_count - 1
            nodes = itertools.chain(job['nodes'],
                        self.object_ring.get_more_nodes(int(job['partition'])))
            while attempts_left > 0:
                # If this throws StopIterator it will be caught way below
                node = next(nodes)
                attempts_left -= 1
                try:
                    with Timeout(self.http_timeout):
                        resp = http_connect(node['ip'], node['port'],
                                node['device'], job['partition'], 'REPLICATE',
                            '', headers={'Content-Length': '0'}).getresponse()
                        if resp.status == 507:
                            self.logger.error(_('%(ip)s/%(device)s responded'
                                    ' as unmounted'), node)
                            attempts_left += 1
                            continue
                        if resp.status != 200:
                            self.logger.error(_("Invalid response %(resp)s "
                                "from %(ip)s"),
                                {'resp': resp.status, 'ip': node['ip']})
                            continue
                        remote_hash = pickle.loads(resp.read())
                        del resp
                    suffixes = [suffix for suffix in local_hash if
                            local_hash[suffix] != remote_hash.get(suffix, -1)]
                    if not suffixes:
                        continue
                    hashed, recalc_hash = tpool.execute(tpooled_get_hashes,
                        job['path'], recalculate=suffixes,
                        reclaim_age=self.reclaim_age)
                    # See tpooled_get_hashes "Hack".
                    if isinstance(hashed, BaseException):
                        raise hashed
                    local_hash = recalc_hash
                    suffixes = [suffix for suffix in local_hash if
                            local_hash[suffix] != remote_hash.get(suffix, -1)]
                    self.rsync(node, job, suffixes)
                    with Timeout(self.http_timeout):
                        conn = http_connect(node['ip'], node['port'],
                            node['device'], job['partition'], 'REPLICATE',
                            '/' + '-'.join(suffixes),
                            headers={'Content-Length': '0'})
                        conn.getresponse().read()
                    self.suffix_sync += len(suffixes)
                except (Exception, Timeout):
                    self.logger.exception(_("Error syncing with node: %s") %
                                            node)
            self.suffix_count += len(local_hash)
        except (Exception, Timeout):
            self.logger.exception(_("Error syncing partition"))
        finally:
            self.partition_times.append(time.time() - begin)

    def stats_line(self):
        """
Logs various stats for the currently running replication pass.
"""
        if self.replication_count:
            rate = self.replication_count / (time.time() - self.start)
            self.logger.info(_("%(replicated)d/%(total)d (%(percentage).2f%%)"
                " partitions replicated in %(time).2fs (%(rate).2f/sec, "
                "%(remaining)s remaining)"),
                {'replicated': self.replication_count, 'total': self.job_count,
                 'percentage': self.replication_count * 100.0 / self.job_count,
                 'time': time.time() - self.start, 'rate': rate,
                 'remaining': '%d%s' % compute_eta(self.start,
                           self.replication_count, self.job_count)})
            if self.suffix_count:
                self.logger.info(_("%(checked)d suffixes checked - "
                    "%(hashed).2f%% hashed, %(synced).2f%% synced"),
                    {'checked': self.suffix_count,
                     'hashed': (self.suffix_hash * 100.0) / self.suffix_count,
                     'synced': (self.suffix_sync * 100.0) / self.suffix_count})
                self.partition_times.sort()
                self.logger.info(_("Partition times: max %(max).4fs, "
                    "min %(min).4fs, med %(med).4fs"),
                    {'max': self.partition_times[-1],
                     'min': self.partition_times[0],
                     'med': self.partition_times[
                                len(self.partition_times) // 2]})
        else:
            self.logger.info(_("Nothing replicated for %s seconds."),
                (time.time() - self.start))

    def kill_coros(self):
        """Utility function that kills all coroutines currently running."""
        for coro in list(self.run_pool.coroutines_running):
            try:
                coro.kill(GreenletExit)
            except GreenletExit:
                pass

    def heartbeat(self):
        """
Loop that runs in the background during replication. It periodically
logs progress.
"""
        while True:
            eventlet.sleep(self.stats_interval)
            self.stats_line()

    def detect_lockups(self):
        """
In testing, the pool.waitall() call very occasionally failed to return.
This is an attempt to make sure the replicator finishes its replication
pass in some eventuality.
"""
        while True:
            eventlet.sleep(self.lockup_timeout)
            if self.replication_count == self.last_replication_count:
                self.logger.error(_("Lockup detected.. killing live coros."))
                self.kill_coros()
            self.last_replication_count = self.replication_count

    def collect_jobs(self):
        """
Returns a sorted list of jobs (dictionaries) that specify the
partitions, nodes, etc to be rsynced.
"""
        jobs = []
        ips = whataremyips()
        for local_dev in [dev for dev in self.object_ring.devs
                if dev and dev['ip'] in ips and dev['port'] == self.port]:
            dev_path = join(self.devices_dir, local_dev['device'])
            obj_path = join(dev_path, 'objects')
            tmp_path = join(dev_path, 'tmp')
            if self.mount_check and not os.path.ismount(dev_path):
                self.logger.warn(_('%s is not mounted'), local_dev['device'])
                continue
            unlink_older_than(tmp_path, time.time() - self.reclaim_age)
            if not os.path.exists(obj_path):
                continue
            for partition in os.listdir(obj_path):
                try:
                    nodes = [node for node in
                        self.object_ring.get_part_nodes(int(partition))
                             if node['id'] != local_dev['id']]
                    jobs.append(dict(path=join(obj_path, partition),
                        nodes=nodes,
                        delete=len(nodes) > self.object_ring.replica_count - 1,
                        partition=partition))
                except ValueError:
                    continue
        random.shuffle(jobs)
        # Partititons that need to be deleted take priority
        jobs.sort(key=lambda job: not job['delete'])
        self.job_count = len(jobs)
        return jobs

    def replicate(self):
        """Run a replication pass"""
        self.start = time.time()
        self.suffix_count = 0
        self.suffix_sync = 0
        self.suffix_hash = 0
        self.replication_count = 0
        self.last_replication_count = -1
        self.partition_times = []
        stats = eventlet.spawn(self.heartbeat)
        lockup_detector = eventlet.spawn(self.detect_lockups)
        eventlet.sleep() # Give spawns a cycle
        try:
            self.run_pool = GreenPool(size=self.concurrency)
            jobs = self.collect_jobs()
            for job in jobs:
                if not self.check_ring():
                    self.logger.info(_("Ring change detected. Aborting "
                            "current replication pass."))
                    return
                if job['delete']:
                    self.run_pool.spawn(self.update_deleted, job)
                else:
                    self.run_pool.spawn(self.update, job)
            with Timeout(self.lockup_timeout):
                self.run_pool.waitall()
        except (Exception, Timeout):
            self.logger.exception(_("Exception in top-level replication loop"))
            self.kill_coros()
        finally:
            stats.kill()
            lockup_detector.kill()
            self.stats_line()

    def run_once(self, *args, **kwargs):
        start = time.time()
        self.logger.info(_("Running object replicator in script mode."))
        self.replicate()
        total = (time.time() - start) / 60
        self.logger.info(
            _("Object replication complete. (%.02f minutes)"), total)
        if self.recon_enable:
            try:
                dump_recon_cache('object_replication_time', total, \
                    self.recon_object)
            except (Exception, Timeout):
                self.logger.exception(_('Exception dumping recon cache'))

    def run_forever(self, *args, **kwargs):
        self.logger.info(_("Starting object replicator in daemon mode."))
        # Run the replicator continually
        while True:
            start = time.time()
            self.logger.info(_("Starting object replication pass."))
            # Run the replicator
            self.replicate()
            total = (time.time() - start) / 60
            self.logger.info(
                _("Object replication complete. (%.02f minutes)"), total)
            if self.recon_enable:
                try:
                    dump_recon_cache('object_replication_time', total, \
                        self.recon_object)
                except (Exception, Timeout):
                    self.logger.exception(_('Exception dumping recon cache'))
            self.logger.debug(_('Replication sleeping for %s seconds.'),
                self.run_pause)
            sleep(self.run_pause)
Something went wrong with that request. Please try again.