1
# Copyright (C) 2006 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""On-disk mutex protecting a resource
19
bzr on-disk objects are locked by the existence of a directory with a
20
particular name within the control directory. We use this rather than OS
21
internal locks (such as flock etc) because they can be seen across all
22
transports, including http.
24
Objects can be read if there is only physical read access; therefore
25
readers can never be required to create a lock, though they will
26
check whether a writer is using the lock. Writers can't detect
27
whether anyone else is reading from the resource as they write.
28
This works because of ordering constraints that make sure readers
29
see a consistent view of existing data.
31
Waiting for a lock must be done by polling; this can be aborted after
34
Locks must always be explicitly released, typically from a try/finally
35
block -- they are not released from a finalizer or when Python
38
Locks may fail to be released if the process is abruptly terminated
39
(machine stop, SIGKILL) or if a remote transport becomes permanently
40
disconnected. There is therefore a method to break an existing lock.
41
This should rarely be used, and generally only with user approval.
42
Locks contain some information on when the lock was taken and by who
43
which may guide in deciding whether it can safely be broken. (This is
44
similar to the messages displayed by emacs and vim.) Note that if the
45
lock holder is still alive they will get no notification that the lock
46
has been broken and will continue their work -- so it is important to be
47
sure they are actually dead.
49
A lock is represented on disk by a directory of a particular name,
50
containing an information file. Taking a lock is done by renaming a
51
temporary directory into place. We use temporary directories because
52
for all known transports and filesystems we believe that exactly one
53
attempt to claim the lock will succeed and the others will fail. (Files
54
won't do because some filesystems or transports only have
55
rename-and-overwrite, making it hard to tell who won.)
57
The desired characteristics are:
59
* Locks are not reentrant. (That is, a client that tries to take a
60
lock it already holds may deadlock or fail.)
61
* Stale locks can be guessed at by a heuristic
62
* Lost locks can be broken by any client
63
* Failed lock operations leave little or no mess
64
* Deadlocks are avoided by having a timeout always in use, clients
65
desiring indefinite waits can retry or set a silly big timeout.
67
Storage formats use the locks, and also need to consider concurrency
68
issues underneath the lock. A format may choose not to use a lock
69
at all for some operations.
71
LockDirs always operate over a Transport. The transport may be readonly, in
72
which case the lock can be queried but not acquired.
74
Locks are identified by a path name, relative to a base transport.
76
Calling code will typically want to make sure there is exactly one LockDir
77
object per actual lock on disk. This module does nothing to prevent aliasing
78
and deadlocks will likely occur if the locks are aliased.
80
In the future we may add a "freshen" method which can be called
81
by a lock holder to check that their lock has not been broken, and to
82
update the timestamp within it.
86
>>> from bzrlib.transport.memory import MemoryTransport
87
>>> # typically will be obtained from a BzrDir, Branch, etc
88
>>> t = MemoryTransport()
89
>>> l = LockDir(t, 'sample-lock')
92
>>> # do something here
99
from cStringIO import StringIO
102
from bzrlib.errors import (
114
from bzrlib.trace import mutter, note
115
from bzrlib.transport import Transport
116
from bzrlib.osutils import rand_chars, format_delta
117
from bzrlib.rio import read_stanza, Stanza
120
# XXX: At the moment there is no consideration of thread safety on LockDir
121
# objects. This should perhaps be updated - e.g. if two threads try to take a
122
# lock at the same time they should *both* get it. But then that's unlikely
125
# TODO: Perhaps store some kind of note like the bzr command line in the lock
128
# TODO: Some kind of callback run while polling a lock to show progress
131
# TODO: Make sure to pass the right file and directory mode bits to all
132
# files/dirs created.
135
_DEFAULT_TIMEOUT_SECONDS = 300
136
_DEFAULT_POLL_SECONDS = 1.0
139
class LockDir(object):
140
"""Write-lock guarding access to data."""
142
__INFO_NAME = '/info'
144
def __init__(self, transport, path, file_modebits=0644, dir_modebits=0755):
145
"""Create a new LockDir object.
147
The LockDir is initially unlocked - this just creates the object.
149
:param transport: Transport which will contain the lock
151
:param path: Path to the lock within the base directory of the
154
assert isinstance(transport, Transport), \
155
("not a transport: %r" % transport)
156
self.transport = transport
158
self._lock_held = False
159
self._fake_read_lock = False
160
self._held_dir = path + '/held'
161
self._held_info_path = self._held_dir + self.__INFO_NAME
162
self._file_modebits = file_modebits
163
self._dir_modebits = dir_modebits
164
self.nonce = rand_chars(20)
166
self._report_function = note
169
return '%s(%s%s)' % (self.__class__.__name__,
173
is_held = property(lambda self: self._lock_held)
175
def create(self, mode=None):
176
"""Create the on-disk lock.
178
This is typically only called when the object/directory containing the
179
directory is first created. The lock is not held when it's created.
181
if self.transport.is_readonly():
182
raise UnlockableTransport(self.transport)
183
self.transport.mkdir(self.path, mode=mode)
185
def attempt_lock(self):
186
"""Take the lock; fail if it's already held.
188
If you wish to block until the lock can be obtained, call wait_lock()
191
if self._fake_read_lock:
192
raise LockContention(self)
193
if self.transport.is_readonly():
194
raise UnlockableTransport(self.transport)
196
tmpname = '%s/pending.%s.tmp' % (self.path, rand_chars(20))
198
self.transport.mkdir(tmpname)
200
# This may raise a FileExists exception
201
# which is okay, it will be caught later and determined
202
# to be a LockContention.
203
self.create(mode=self._dir_modebits)
205
# After creating the lock directory, try again
206
self.transport.mkdir(tmpname)
208
info_bytes = self._prepare_info()
209
# We use put_file_non_atomic because we just created a new unique
210
# directory so we don't have to worry about files existing there.
211
# We'll rename the whole directory into place to get atomic
213
self.transport.put_bytes_non_atomic(tmpname + self.__INFO_NAME,
216
self.transport.rename(tmpname, self._held_dir)
217
self._lock_held = True
219
except (PathError, DirectoryNotEmpty, FileExists, ResourceBusy), e:
220
mutter("contention on %r: %s", self, e)
221
raise LockContention(self)
224
"""Release a held lock
226
if self._fake_read_lock:
227
self._fake_read_lock = False
229
if not self._lock_held:
230
raise LockNotHeld(self)
231
# rename before deleting, because we can't atomically remove the whole
233
tmpname = '%s/releasing.%s.tmp' % (self.path, rand_chars(20))
234
# gotta own it to unlock
236
self.transport.rename(self._held_dir, tmpname)
237
self._lock_held = False
238
self.transport.delete(tmpname + self.__INFO_NAME)
239
self.transport.rmdir(tmpname)
241
def break_lock(self):
242
"""Break a lock not held by this instance of LockDir.
244
This is a UI centric function: it uses the bzrlib.ui.ui_factory to
245
prompt for input if a lock is detected and there is any doubt about
246
it possibly being still active.
248
self._check_not_locked()
249
holder_info = self.peek()
250
if holder_info is not None:
251
lock_info = '\n'.join(self._format_lock_info(holder_info))
252
if bzrlib.ui.ui_factory.get_boolean("Break %s" % lock_info):
253
self.force_break(holder_info)
255
def force_break(self, dead_holder_info):
256
"""Release a lock held by another process.
258
WARNING: This should only be used when the other process is dead; if
259
it still thinks it has the lock there will be two concurrent writers.
260
In general the user's approval should be sought for lock breaks.
262
dead_holder_info must be the result of a previous LockDir.peek() call;
263
this is used to check that it's still held by the same process that
264
the user decided was dead. If this is not the current holder,
265
LockBreakMismatch is raised.
267
After the lock is broken it will not be held by any process.
268
It is possible that another process may sneak in and take the
269
lock before the breaking process acquires it.
271
if not isinstance(dead_holder_info, dict):
272
raise ValueError("dead_holder_info: %r" % dead_holder_info)
273
self._check_not_locked()
274
current_info = self.peek()
275
if current_info is None:
276
# must have been recently released
278
if current_info != dead_holder_info:
279
raise LockBreakMismatch(self, current_info, dead_holder_info)
280
tmpname = '%s/broken.%s.tmp' % (self.path, rand_chars(20))
281
self.transport.rename(self._held_dir, tmpname)
282
# check that we actually broke the right lock, not someone else;
283
# there's a small race window between checking it and doing the
285
broken_info_path = tmpname + self.__INFO_NAME
286
broken_info = self._read_info_file(broken_info_path)
287
if broken_info != dead_holder_info:
288
raise LockBreakMismatch(self, broken_info, dead_holder_info)
289
self.transport.delete(broken_info_path)
290
self.transport.rmdir(tmpname)
292
def _check_not_locked(self):
293
"""If the lock is held by this instance, raise an error."""
295
raise AssertionError("can't break own lock: %r" % self)
298
"""Make sure that the lock is still held by this locker.
300
This should only fail if the lock was broken by user intervention,
301
or if the lock has been affected by a bug.
303
If the lock is not thought to be held, raises LockNotHeld. If
304
the lock is thought to be held but has been broken, raises
307
if not self._lock_held:
308
raise LockNotHeld(self)
311
# no lock there anymore!
312
raise LockBroken(self)
313
if info.get('nonce') != self.nonce:
314
# there is a lock, but not ours
315
raise LockBroken(self)
317
def _read_info_file(self, path):
318
"""Read one given info file.
320
peek() reads the info file of the lock holder, if any.
322
return self._parse_info(self.transport.get(path))
325
"""Check if the lock is held by anyone.
327
If it is held, this returns the lock info structure as a rio Stanza,
328
which contains some information about the current lock holder.
329
Otherwise returns None.
332
info = self._read_info_file(self._held_info_path)
333
assert isinstance(info, dict), \
334
"bad parse result %r" % info
336
except NoSuchFile, e:
339
def _prepare_info(self):
340
"""Write information about a pending lock to a temporary file.
343
# XXX: is creating this here inefficient?
344
config = bzrlib.config.GlobalConfig()
345
s = Stanza(hostname=socket.gethostname(),
346
pid=str(os.getpid()),
347
start_time=str(int(time.time())),
349
user=config.user_email(),
353
def _parse_info(self, info_file):
354
return read_stanza(info_file.readlines()).as_dict()
356
def wait_lock(self, timeout=None, poll=None):
357
"""Wait a certain period for a lock.
359
If the lock can be acquired within the bounded time, it
360
is taken and this returns. Otherwise, LockContention
361
is raised. Either way, this function should return within
362
approximately `timeout` seconds. (It may be a bit more if
363
a transport operation takes a long time to complete.)
366
timeout = _DEFAULT_TIMEOUT_SECONDS
368
poll = _DEFAULT_POLL_SECONDS
370
# XXX: the transport interface doesn't let us guard
371
# against operations there taking a long time.
372
deadline = time.time() + timeout
379
except LockContention:
381
new_info = self.peek()
382
mutter('last_info: %s, new info: %s', last_info, new_info)
383
if new_info is not None and new_info != last_info:
384
if last_info is None:
385
start = 'Unable to obtain'
387
start = 'Lock owner changed for'
389
formatted_info = self._format_lock_info(new_info)
390
if deadline_str is None:
391
deadline_str = time.strftime('%H:%M:%S',
392
time.localtime(deadline))
393
self._report_function('%s %s\n'
395
'%s\n' # locked ... ago
396
'Will continue to try until %s\n',
403
if time.time() + poll < deadline:
406
raise LockContention(self)
408
def lock_write(self):
409
"""Wait for and acquire the lock."""
413
"""Compatibility-mode shared lock.
415
LockDir doesn't support shared read-only locks, so this
416
just pretends that the lock is taken but really does nothing.
418
# At the moment Branches are commonly locked for read, but
419
# we can't rely on that remotely. Once this is cleaned up,
420
# reenable this warning to prevent it coming back in
422
## warn("LockDir.lock_read falls back to write lock")
423
if self._lock_held or self._fake_read_lock:
424
raise LockContention(self)
425
self._fake_read_lock = True
427
def wait(self, timeout=20, poll=0.5):
428
"""Wait a certain period for a lock to be released."""
429
# XXX: the transport interface doesn't let us guard
430
# against operations there taking a long time.
431
deadline = time.time() + timeout
435
if time.time() + poll < deadline:
438
raise LockContention(self)
440
def _format_lock_info(self, info):
441
"""Turn the contents of peek() into something for the user"""
442
lock_url = self.transport.abspath(self.path)
443
delta = time.time() - int(info['start_time'])
445
'lock %s' % (lock_url,),
446
'held by %(user)s on host %(hostname)s [process #%(pid)s]' % info,
447
'locked %s' % (format_delta(delta),),
added
removed
bzrlib/lockdir.py
