tahoe-lafs/src/allmydata/mutable/node.py

477 lines
19 KiB
Python

import weakref, random
from twisted.application import service
from zope.interface import implements
from twisted.internet import defer, reactor
from foolscap.eventual import eventually
from allmydata.interfaces import IMutableFileNode, IMutableFileURI, \
ICheckable, ICheckerResults
from allmydata.util import hashutil, log
from allmydata.util.assertutil import precondition
from allmydata.uri import WriteableSSKFileURI
from allmydata.immutable.encode import NotEnoughSharesError
from pycryptopp.publickey import rsa
from pycryptopp.cipher.aes import AES
from publish import Publish
from common import MODE_READ, MODE_WRITE, UnrecoverableFileError, \
ResponseCache, UncoordinatedWriteError
from servermap import ServerMap, ServermapUpdater
from retrieve import Retrieve
from checker import MutableChecker, MutableCheckAndRepairer
from repair import Repairer
class BackoffAgent:
# these parameters are copied from foolscap.reconnector, which gets them
# from twisted.internet.protocol.ReconnectingClientFactory
initialDelay = 1.0
factor = 2.7182818284590451 # (math.e)
jitter = 0.11962656492 # molar Planck constant times c, Joule meter/mole
maxRetries = 4
def __init__(self):
self._delay = self.initialDelay
self._count = 0
def delay(self, node, f):
self._count += 1
if self._count == 4:
return f
self._delay = self._delay * self.factor
self._delay = random.normalvariate(self._delay,
self._delay * self.jitter)
d = defer.Deferred()
reactor.callLater(self._delay, d.callback, None)
return d
# use client.create_mutable_file() to make one of these
class MutableFileNode:
implements(IMutableFileNode, ICheckable)
SIGNATURE_KEY_SIZE = 2048
DEFAULT_ENCODING = (3, 10)
checker_class = MutableChecker
check_and_repairer_class = MutableCheckAndRepairer
def __init__(self, client):
self._client = client
self._pubkey = None # filled in upon first read
self._privkey = None # filled in if we're mutable
# we keep track of the last encoding parameters that we use. These
# are updated upon retrieve, and used by publish. If we publish
# without ever reading (i.e. overwrite()), then we use these values.
(self._required_shares, self._total_shares) = self.DEFAULT_ENCODING
self._sharemap = {} # known shares, shnum-to-[nodeids]
self._cache = ResponseCache()
# all users of this MutableFileNode go through the serializer. This
# takes advantage of the fact that Deferreds discard the callbacks
# that they're done with, so we can keep using the same Deferred
# forever without consuming more and more memory.
self._serializer = defer.succeed(None)
def __repr__(self):
if hasattr(self, '_uri'):
return "<%s %x %s %s>" % (self.__class__.__name__, id(self), self.is_readonly() and 'RO' or 'RW', self._uri.abbrev())
else:
return "<%s %x %s %s>" % (self.__class__.__name__, id(self), None, None)
def init_from_uri(self, myuri):
# we have the URI, but we have not yet retrieved the public
# verification key, nor things like 'k' or 'N'. If and when someone
# wants to get our contents, we'll pull from shares and fill those
# in.
self._uri = IMutableFileURI(myuri)
if not self._uri.is_readonly():
self._writekey = self._uri.writekey
self._readkey = self._uri.readkey
self._storage_index = self._uri.storage_index
self._fingerprint = self._uri.fingerprint
# the following values are learned during Retrieval
# self._pubkey
# self._required_shares
# self._total_shares
# and these are needed for Publish. They are filled in by Retrieval
# if possible, otherwise by the first peer that Publish talks to.
self._privkey = None
self._encprivkey = None
return self
def create(self, initial_contents, keypair_generator=None):
"""Call this when the filenode is first created. This will generate
the keys, generate the initial shares, wait until at least numpeers
are connected, allocate shares, and upload the initial
contents. Returns a Deferred that fires (with the MutableFileNode
instance you should use) when it completes.
"""
d = defer.maybeDeferred(self._generate_pubprivkeys, keypair_generator)
d.addCallback(self._generated)
d.addCallback(lambda res: self._upload(initial_contents, None))
return d
def _generated(self, (pubkey, privkey) ):
self._pubkey, self._privkey = pubkey, privkey
pubkey_s = self._pubkey.serialize()
privkey_s = self._privkey.serialize()
self._writekey = hashutil.ssk_writekey_hash(privkey_s)
self._encprivkey = self._encrypt_privkey(self._writekey, privkey_s)
self._fingerprint = hashutil.ssk_pubkey_fingerprint_hash(pubkey_s)
self._uri = WriteableSSKFileURI(self._writekey, self._fingerprint)
self._readkey = self._uri.readkey
self._storage_index = self._uri.storage_index
self._required_shares, self._total_shares = self.DEFAULT_ENCODING
def _generate_pubprivkeys(self, keypair_generator):
if keypair_generator:
return keypair_generator(self.SIGNATURE_KEY_SIZE)
else:
# RSA key generation for a 2048 bit key takes between 0.8 and 3.2 secs
signer = rsa.generate(self.SIGNATURE_KEY_SIZE)
verifier = signer.get_verifying_key()
return verifier, signer
def _encrypt_privkey(self, writekey, privkey):
enc = AES(writekey)
crypttext = enc.process(privkey)
return crypttext
def _decrypt_privkey(self, enc_privkey):
enc = AES(self._writekey)
privkey = enc.process(enc_privkey)
return privkey
def _populate_pubkey(self, pubkey):
self._pubkey = pubkey
def _populate_required_shares(self, required_shares):
self._required_shares = required_shares
def _populate_total_shares(self, total_shares):
self._total_shares = total_shares
def _populate_privkey(self, privkey):
self._privkey = privkey
def _populate_encprivkey(self, encprivkey):
self._encprivkey = encprivkey
def get_write_enabler(self, peerid):
assert len(peerid) == 20
return hashutil.ssk_write_enabler_hash(self._writekey, peerid)
def get_renewal_secret(self, peerid):
assert len(peerid) == 20
crs = self._client.get_renewal_secret()
frs = hashutil.file_renewal_secret_hash(crs, self._storage_index)
return hashutil.bucket_renewal_secret_hash(frs, peerid)
def get_cancel_secret(self, peerid):
assert len(peerid) == 20
ccs = self._client.get_cancel_secret()
fcs = hashutil.file_cancel_secret_hash(ccs, self._storage_index)
return hashutil.bucket_cancel_secret_hash(fcs, peerid)
def get_writekey(self):
return self._writekey
def get_readkey(self):
return self._readkey
def get_storage_index(self):
return self._storage_index
def get_privkey(self):
return self._privkey
def get_encprivkey(self):
return self._encprivkey
def get_pubkey(self):
return self._pubkey
def get_required_shares(self):
return self._required_shares
def get_total_shares(self):
return self._total_shares
####################################
# IFilesystemNode
def get_uri(self):
return self._uri.to_string()
def get_size(self):
return "?" # TODO: this is likely to cause problems, not being an int
def get_readonly(self):
if self.is_readonly():
return self
ro = MutableFileNode(self._client)
ro.init_from_uri(self._uri.get_readonly())
return ro
def get_readonly_uri(self):
return self._uri.get_readonly().to_string()
def is_mutable(self):
return self._uri.is_mutable()
def is_readonly(self):
return self._uri.is_readonly()
def __hash__(self):
return hash((self.__class__, self._uri))
def __cmp__(self, them):
if cmp(type(self), type(them)):
return cmp(type(self), type(them))
if cmp(self.__class__, them.__class__):
return cmp(self.__class__, them.__class__)
return cmp(self._uri, them._uri)
def get_verifier(self):
return IMutableFileURI(self._uri).get_verifier()
def _do_serialized(self, cb, *args, **kwargs):
# note: to avoid deadlock, this callable is *not* allowed to invoke
# other serialized methods within this (or any other)
# MutableFileNode. The callable should be a bound method of this same
# MFN instance.
d = defer.Deferred()
self._serializer.addCallback(lambda ignore: cb(*args, **kwargs))
# we need to put off d.callback until this Deferred is finished being
# processed. Otherwise the caller's subsequent activities (like,
# doing other things with this node) can cause reentrancy problems in
# the Deferred code itself
self._serializer.addBoth(lambda res: eventually(d.callback, res))
# add a log.err just in case something really weird happens, because
# self._serializer stays around forever, therefore we won't see the
# usual Unhandled Error in Deferred that would give us a hint.
self._serializer.addErrback(log.err)
return d
#################################
# ICheckable
def check(self, verify=False):
checker = self.checker_class(self)
return checker.check(verify)
def check_and_repair(self, verify=False):
checker = self.check_and_repairer_class(self)
return checker.check(verify)
#################################
# IRepairable
def repair(self, checker_results, force=False):
assert ICheckerResults(checker_results)
r = Repairer(self, checker_results)
d = r.start(force)
return d
#################################
# IMutableFileNode
# allow the use of IDownloadTarget
def download(self, target):
# fake it. TODO: make this cleaner.
d = self.download_best_version()
def _done(data):
target.open(len(data))
target.write(data)
target.close()
return target.finish()
d.addCallback(_done)
return d
# new API
def download_best_version(self):
return self._do_serialized(self._download_best_version)
def _download_best_version(self):
servermap = ServerMap()
d = self._try_once_to_download_best_version(servermap, MODE_READ)
def _maybe_retry(f):
f.trap(NotEnoughSharesError)
# the download is worth retrying once. Make sure to use the
# old servermap, since it is what remembers the bad shares,
# but use MODE_WRITE to make it look for even more shares.
# TODO: consider allowing this to retry multiple times.. this
# approach will let us tolerate about 8 bad shares, I think.
return self._try_once_to_download_best_version(servermap,
MODE_WRITE)
d.addErrback(_maybe_retry)
return d
def _try_once_to_download_best_version(self, servermap, mode):
d = self._update_servermap(servermap, mode)
d.addCallback(self._once_updated_download_best_version, servermap)
return d
def _once_updated_download_best_version(self, ignored, servermap):
goal = servermap.best_recoverable_version()
if not goal:
raise UnrecoverableFileError("no recoverable versions")
return self._try_once_to_download_version(servermap, goal)
def get_size_of_best_version(self):
d = self.get_servermap(MODE_READ)
def _got_servermap(smap):
ver = smap.best_recoverable_version()
if not ver:
raise UnrecoverableFileError("no recoverable version")
return smap.size_of_version(ver)
d.addCallback(_got_servermap)
return d
def overwrite(self, new_contents):
return self._do_serialized(self._overwrite, new_contents)
def _overwrite(self, new_contents):
servermap = ServerMap()
d = self._update_servermap(servermap, mode=MODE_WRITE)
d.addCallback(lambda ignored: self._upload(new_contents, servermap))
return d
def modify(self, modifier, backoffer=None):
"""I use a modifier callback to apply a change to the mutable file.
I implement the following pseudocode::
obtain_mutable_filenode_lock()
while True:
update_servermap(MODE_WRITE)
old = retrieve_best_version()
new = modifier(old, *args, **kwargs)
if new == old: break
try:
publish(new)
except UncoordinatedWriteError, e:
backoffer(e)
continue
break
release_mutable_filenode_lock()
The idea is that your modifier function can apply a delta of some
sort, and it will be re-run as necessary until it succeeds. The
modifier must inspect the old version to see whether its delta has
already been applied: if so it should return the contents unmodified.
Note that the modifier is required to run synchronously, and must not
invoke any methods on this MutableFileNode instance.
The backoff-er is a callable that is responsible for inserting a
random delay between subsequent attempts, to help competing updates
from colliding forever. It is also allowed to give up after a while.
The backoffer is given two arguments: this MutableFileNode, and the
Failure object that contains the UncoordinatedWriteError. It should
return a Deferred that will fire when the next attempt should be
made, or return the Failure if the loop should give up. If
backoffer=None, a default one is provided which will perform
exponential backoff, and give up after 4 tries. Note that the
backoffer should not invoke any methods on this MutableFileNode
instance, and it needs to be highly conscious of deadlock issues.
"""
return self._do_serialized(self._modify, modifier, backoffer)
def _modify(self, modifier, backoffer):
servermap = ServerMap()
if backoffer is None:
backoffer = BackoffAgent().delay
return self._modify_and_retry(servermap, modifier, backoffer)
def _modify_and_retry(self, servermap, modifier, backoffer):
d = self._modify_once(servermap, modifier)
def _retry(f):
f.trap(UncoordinatedWriteError)
d2 = defer.maybeDeferred(backoffer, self, f)
d2.addCallback(lambda ignored:
self._modify_and_retry(servermap, modifier,
backoffer))
return d2
d.addErrback(_retry)
return d
def _modify_once(self, servermap, modifier):
d = self._update_servermap(servermap, MODE_WRITE)
d.addCallback(self._once_updated_download_best_version, servermap)
def _apply(old_contents):
new_contents = modifier(old_contents)
if new_contents is None or new_contents == old_contents:
# no changes need to be made
return
precondition(isinstance(new_contents, str),
"Modifier function must return a string or None")
return self._upload(new_contents, servermap)
d.addCallback(_apply)
return d
def get_servermap(self, mode):
return self._do_serialized(self._get_servermap, mode)
def _get_servermap(self, mode):
servermap = ServerMap()
return self._update_servermap(servermap, mode)
def _update_servermap(self, servermap, mode):
u = ServermapUpdater(self, servermap, mode)
self._client.notify_mapupdate(u.get_status())
return u.update()
def download_version(self, servermap, version, fetch_privkey=False):
return self._do_serialized(self._try_once_to_download_version,
servermap, version, fetch_privkey)
def _try_once_to_download_version(self, servermap, version,
fetch_privkey=False):
r = Retrieve(self, servermap, version, fetch_privkey)
self._client.notify_retrieve(r.get_status())
return r.download()
def upload(self, new_contents, servermap):
return self._do_serialized(self._upload, new_contents, servermap)
def _upload(self, new_contents, servermap):
assert self._pubkey, "update_servermap must be called before publish"
p = Publish(self, servermap)
self._client.notify_publish(p.get_status(), len(new_contents))
return p.publish(new_contents)
class MutableWatcher(service.MultiService):
MAX_MAPUPDATE_STATUSES = 20
MAX_PUBLISH_STATUSES = 20
MAX_RETRIEVE_STATUSES = 20
name = "mutable-watcher"
def __init__(self, stats_provider=None):
service.MultiService.__init__(self)
self.stats_provider = stats_provider
self._all_mapupdate_status = weakref.WeakKeyDictionary()
self._recent_mapupdate_status = []
self._all_publish_status = weakref.WeakKeyDictionary()
self._recent_publish_status = []
self._all_retrieve_status = weakref.WeakKeyDictionary()
self._recent_retrieve_status = []
def notify_mapupdate(self, p):
self._all_mapupdate_status[p] = None
self._recent_mapupdate_status.append(p)
while len(self._recent_mapupdate_status) > self.MAX_MAPUPDATE_STATUSES:
self._recent_mapupdate_status.pop(0)
def notify_publish(self, p, size):
self._all_publish_status[p] = None
self._recent_publish_status.append(p)
if self.stats_provider:
self.stats_provider.count('mutable.files_published', 1)
# We must be told bytes_published as an argument, since the
# publish_status does not yet know how much data it will be asked
# to send. When we move to MDMF we'll need to find a better way
# to handle this.
self.stats_provider.count('mutable.bytes_published', size)
while len(self._recent_publish_status) > self.MAX_PUBLISH_STATUSES:
self._recent_publish_status.pop(0)
def notify_retrieve(self, r):
self._all_retrieve_status[r] = None
self._recent_retrieve_status.append(r)
if self.stats_provider:
self.stats_provider.count('mutable.files_retrieved', 1)
self.stats_provider.count('mutable.bytes_retrieved', r.get_size())
while len(self._recent_retrieve_status) > self.MAX_RETRIEVE_STATUSES:
self._recent_retrieve_status.pop(0)
def list_all_mapupdate_statuses(self):
return self._all_mapupdate_status.keys()
def list_all_publish_statuses(self):
return self._all_publish_status.keys()
def list_all_retrieve_statuses(self):
return self._all_retrieve_status.keys()