summaryrefslogtreecommitdiff
path: root/src/leap/soledad/server/_blobs/fs_backend.py
blob: 769ba47bee20f8e7a1ea0bd3ae89b54a200c2dd4 (plain)
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
# -*- coding: utf-8 -*-
# _blobs/fs_backend.py
# Copyright (C) 2017 LEAP
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
"""
A backend for blobs that stores in filesystem.
"""
import base64
import json
import os
import time

from collections import defaultdict
from zope.interface import implementer

from twisted.internet import defer
from twisted.internet import utils
from twisted.web.static import NoRangeStaticProducer
from twisted.web.static import SingleRangeStaticProducer

from leap.common.files import mkdir_p
from leap.soledad.common.blobs import ACCEPTED_FLAGS
from leap.soledad.common.blobs import InvalidFlag
from leap.soledad.common.log import getLogger
from leap.soledad.server import interfaces

from .errors import BlobExists
from .errors import BlobNotFound
from .errors import QuotaExceeded
from .util import VALID_STRINGS


logger = getLogger(__name__)


class NoRangeProducer(NoRangeStaticProducer):
    """
    A static file producer that fires a deferred when it's finished.
    """

    def start(self):
        NoRangeStaticProducer.start(self)
        if self.request is None:
            return defer.succeed(None)
        self.deferred = defer.Deferred()
        return self.deferred

    def stopProducing(self):
        NoRangeStaticProducer.stopProducing(self)
        if hasattr(self, 'deferred'):
            self.deferred.callback(None)


class SingleRangeProducer(SingleRangeStaticProducer):
    """
    A static file producer of a single file range that fires a deferred when
    it's finished.
    """

    def start(self):
        SingleRangeStaticProducer.start(self)
        if self.request is None:
            return defer.succeed(None)
        self.deferred = defer.Deferred()
        return self.deferred

    def stopProducing(self):
        SingleRangeStaticProducer.stopProducing(self)
        if hasattr(self, 'deferred'):
            self.deferred.callback(None)


@implementer(interfaces.IBlobsBackend)
class FilesystemBlobsBackend(object):

    USAGE_TIMEOUT = 30

    def __init__(self, blobs_path='/tmp/blobs/', quota=200 * 1024,
                 concurrent_writes=50):
        self.quota = quota
        self.semaphore = defer.DeferredSemaphore(concurrent_writes)
        if not os.path.isdir(blobs_path):
            os.makedirs(blobs_path)
        self.path = blobs_path
        self.usage = defaultdict(lambda: (None, None))
        self.usage_locks = defaultdict(defer.DeferredLock)

    def __touch(self, path):
        open(path, 'a')

    def _fslock(self, path):
        dirname, _ = os.path.split(path)
        mkdir_p(dirname)
        name = path + '.lock'
        # TODO: evaluate the need to replace this for a readers-writer lock.
        return defer.DeferredFilesystemLock(name)

    @defer.inlineCallbacks
    def read_blob(self, user, blob_id, consumer, namespace='', range=None):
        path = self._get_path(user, blob_id, namespace)
        if not os.path.isfile(path):
            raise BlobNotFound((user, blob_id))
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()

            logger.info('reading blob: %s - %s@%s'
                        % (user, blob_id, namespace))
            logger.debug('blob path: %s' % path)
            with open(path) as fd:
                if range is None:
                    producer = NoRangeProducer(consumer, fd)
                else:
                    start, end = range
                    offset = start
                    size = end - start
                    args = (consumer, fd, offset, size)
                    producer = SingleRangeProducer(*args)
                yield producer.start()
        finally:
            fslock.unlock()

    @defer.inlineCallbacks
    def get_flags(self, user, blob_id, namespace=''):
        path = self._get_path(user, blob_id, namespace)
        if not os.path.isfile(path):
            raise BlobNotFound((user, blob_id))
        if not os.path.isfile(path + '.flags'):
            defer.returnValue([])
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()

            with open(path + '.flags', 'r') as flags_file:
                flags = json.loads(flags_file.read())
                defer.returnValue(flags)
        finally:
            fslock.unlock()

    @defer.inlineCallbacks
    def set_flags(self, user, blob_id, flags, namespace=''):
        path = self._get_path(user, blob_id, namespace)
        if not os.path.isfile(path):
            raise BlobNotFound((user, blob_id))
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()
            for flag in flags:
                if flag not in ACCEPTED_FLAGS:
                    raise InvalidFlag(flag)
            with open(path + '.flags', 'w') as flags_file:
                raw_flags = json.dumps(flags)
                flags_file.write(raw_flags)
        finally:
            fslock.unlock()

    @defer.inlineCallbacks
    def write_blob(self, user, blob_id, producer, namespace=''):
        path = self._get_path(user, blob_id, namespace)
        if os.path.isfile(path):
            raise BlobExists
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()

            try:
                # limit the number of concurrent writes to disk
                yield self.semaphore.acquire()

                try:
                    mkdir_p(os.path.split(path)[0])
                except OSError as e:
                    logger.warn(
                        "Got exception trying to create directory: %r" % e)
                used = yield self.get_total_storage(user)
                length = producer.length / 1024.0
                if used + length > self.quota:
                    raise QuotaExceeded
                logger.info('writing blob: %s - %s' % (user, blob_id))
                with open(path, 'wb') as blobfile:
                    yield producer.startProducing(blobfile)
                used += length
                yield self._update_usage(user, used)
            finally:
                self.semaphore.release()
        finally:
            fslock.unlock()

    @defer.inlineCallbacks
    def _update_usage(self, user, used):
        lock = self.usage_locks[user]
        yield lock.acquire()
        try:
            _, timestamp = self.usage[user]
            self.usage[user] = (used, timestamp)
        finally:
            lock.release()

    @defer.inlineCallbacks
    def delete_blob(self, user, blob_id, namespace=''):
        path = self._get_path(user, blob_id, namespace)
        if not os.path.isfile(path):
            raise BlobNotFound((user, blob_id))
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()
            self.__touch(path + '.deleted')
            os.unlink(path)
            try:
                os.unlink(path + '.flags')
            except Exception:
                pass
        finally:
            fslock.unlock()

    @defer.inlineCallbacks
    def get_blob_size(self, user, blob_id, namespace=''):
        path = self._get_path(user, blob_id, namespace)
        if not os.path.isfile(path):
            raise BlobNotFound((user, blob_id))
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()
            size = os.stat(path).st_size
            defer.returnValue(size)
        finally:
            fslock.unlock()

    def count(self, user, namespace=''):
        try:
            base_path = self._get_path(user, namespace=namespace)
        except Exception as e:
            return defer.fail(e)
        count = 0
        for _, _, filenames in os.walk(base_path):
            count += len(filter(lambda i: not i.endswith('.flags'), filenames))
        return defer.succeed(count)

    def list_blobs(self, user, namespace='', order_by=None, deleted=False,
                   filter_flag=False):
        namespace = namespace or 'default'
        blob_ids = []
        try:
            base_path = self._get_path(user, namespace=namespace)
        except Exception as e:
            return defer.fail(e)

        def match(name):
            if deleted:
                return name.endswith('.deleted')
            return VALID_STRINGS.match(name)
        for root, dirs, filenames in os.walk(base_path):
            blob_ids += [os.path.join(root, name) for name in filenames
                         if match(name)]
        if order_by in ['date', '+date']:
            blob_ids.sort(key=lambda x: os.path.getmtime(x))
        elif order_by == '-date':
            blob_ids.sort(key=lambda x: os.path.getmtime(x), reverse=True)
        elif order_by:
            exc = Exception("Unsupported order_by parameter: %s" % order_by)
            return defer.fail(exc)
        if filter_flag:
            blob_ids = list(self._filter_flag(blob_ids, filter_flag))
        blob_ids = [os.path.basename(path).replace('.deleted', '')
                    for path in blob_ids]
        return defer.succeed(blob_ids)

    def _filter_flag(self, blob_paths, flag):
        for blob_path in blob_paths:
            flag_path = blob_path + '.flags'
            if not os.path.isfile(flag_path):
                continue
            with open(flag_path, 'r') as flags_file:
                blob_flags = json.loads(flags_file.read())
            if flag in blob_flags:
                yield blob_path

    @defer.inlineCallbacks
    def get_total_storage(self, user):
        lock = self.usage_locks[user]
        yield lock.acquire()
        try:
            used, timestamp = self.usage[user]
            if used is None or time.time() > timestamp + self.USAGE_TIMEOUT:
                path = self._get_path(user)
                used = yield self._get_disk_usage(path)
                self.usage[user] = (used, time.time())
            defer.returnValue(used)
        finally:
            lock.release()

    @defer.inlineCallbacks
    def get_tag(self, user, blob_id, namespace=''):
        path = self._get_path(user, blob_id, namespace)
        if not os.path.isfile(path):
            raise BlobNotFound((user, blob_id))
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()
            with open(path) as doc_file:
                doc_file.seek(-16, 2)
                tag = base64.urlsafe_b64encode(doc_file.read())
                defer.returnValue(tag)
        finally:
            fslock.unlock()

    @defer.inlineCallbacks
    def _get_disk_usage(self, start_path):
        if not os.path.isdir(start_path):
            defer.returnValue(0)
        cmd = ['/usr/bin/du', '-s', '-c', start_path]
        output = yield utils.getProcessOutput(cmd[0], cmd[1:])
        size = output.split()[0]
        defer.returnValue(int(size))

    def _validate_path(self, desired_path, user, blob_id):
        if not VALID_STRINGS.match(user):
            raise Exception("Invalid characters on user: %s" % user)
        if blob_id and not VALID_STRINGS.match(blob_id):
            raise Exception("Invalid characters on blob_id: %s" % blob_id)
        desired_path = os.path.realpath(desired_path)  # expand path references
        root = os.path.realpath(self.path)
        if not desired_path.startswith(root + os.sep + user):
            err = "User %s tried accessing a invalid path: %s" % (user,
                                                                  desired_path)
            raise Exception(err)
        return desired_path

    @defer.inlineCallbacks
    def exists(self, user, blob_id, namespace):
        path = self._get_path(user, blob_id, namespace)
        fslock = self._fslock(path)
        try:
            yield fslock.deferUntilLocked()
            defer.returnValue(os.path.isfile(path))
        finally:
            fslock.unlock()

    def _get_path(self, user, blob_id='', namespace=''):
        parts = [user]
        if blob_id:
            namespace = namespace or 'default'
            parts += self._get_path_parts(blob_id, namespace)
        elif namespace and not blob_id:
            parts += [namespace]  # namespace path
        else:
            pass  # root path
        path = os.path.join(self.path, *parts)
        return self._validate_path(path, user, blob_id)

    def _get_path_parts(self, blob_id, custom):
        if custom and not blob_id:
            return [custom]
        return [custom] + [blob_id[0], blob_id[0:3], blob_id[0:6]] + [blob_id]