ExternalVendorCode/tahoe-lafs
1
0
Fork 0
mirror of https://github.com/tahoe-lafs/tahoe-lafs.git synced 2024-12-21 05:53:12 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
7f9b715b3e
tahoe-lafs/src/allmydata/interfaces.py
Brian Warner d1d988410b remove "key-generator" node type and client support 
closes ticket:2783
2016-05-04 16:53:04 -07:00

2833 lines
123 KiB
Python
Raw Blame History

from zope.interface import Interface, Attribute
from foolscap.api import StringConstraint, ListOf, TupleOf, SetOf, DictOf, \
ChoiceOf, IntegerConstraint, Any, RemoteInterface, Referenceable
HASH_SIZE=32
SALT_SIZE=16
SDMF_VERSION=0
MDMF_VERSION=1
Hash = StringConstraint(maxLength=HASH_SIZE,
minLength=HASH_SIZE)# binary format 32-byte SHA256 hash
Nodeid = StringConstraint(maxLength=20,
minLength=20) # binary format 20-byte SHA1 hash
FURL = StringConstraint(1000)
StorageIndex = StringConstraint(16)
URI = StringConstraint(300) # kind of arbitrary
MAX_BUCKETS = 256 # per peer -- zfec offers at most 256 shares per file
DEFAULT_MAX_SEGMENT_SIZE = 128*1024
ShareData = StringConstraint(None)
URIExtensionData = StringConstraint(1000)
Number = IntegerConstraint(8) # 2**(8*8) == 16EiB ~= 18e18 ~= 18 exabytes
Offset = Number
ReadSize = int # the 'int' constraint is 2**31 == 2Gib -- large files are processed in not-so-large increments
WriteEnablerSecret = Hash # used to protect mutable share modifications
LeaseRenewSecret = Hash # used to protect lease renewal requests
LeaseCancelSecret = Hash # was used to protect lease cancellation requests
class RIBucketWriter(RemoteInterface):
""" Objects of this kind live on the server side. """
def write(offset=Offset, data=ShareData):
return None
def close():
"""
If the data that has been written is incomplete or inconsistent then
the server will throw the data away, else it will store it for future
retrieval.
"""
return None
def abort():
"""Abandon all the data that has been written.
"""
return None
class RIBucketReader(RemoteInterface):
def read(offset=Offset, length=ReadSize):
return ShareData
def advise_corrupt_share(reason=str):
"""Clients who discover hash failures in shares that they have
downloaded from me will use this method to inform me about the
failures. I will record their concern so that my operator can
manually inspect the shares in question. I return None.
This is a wrapper around RIStorageServer.advise_corrupt_share()
that is tied to a specific share, and therefore does not need the
extra share-identifying arguments. Please see that method for full
documentation.
"""
TestVector = ListOf(TupleOf(Offset, ReadSize, str, str))
# elements are (offset, length, operator, specimen)
# operator is one of "lt, le, eq, ne, ge, gt"
# nop always passes and is used to fetch data while writing.
# you should use length==len(specimen) for everything except nop
DataVector = ListOf(TupleOf(Offset, ShareData))
# (offset, data). This limits us to 30 writes of 1MiB each per call
TestAndWriteVectorsForShares = DictOf(int,
TupleOf(TestVector,
DataVector,
ChoiceOf(None, Offset), # new_length
))
ReadVector = ListOf(TupleOf(Offset, ReadSize))
ReadData = ListOf(ShareData)
# returns data[offset:offset+length] for each element of TestVector
class RIStorageServer(RemoteInterface):
__remote_name__ = "RIStorageServer.tahoe.allmydata.com"
def get_version():
"""
Return a dictionary of version information.
"""
return DictOf(str, Any())
def allocate_buckets(storage_index=StorageIndex,
renew_secret=LeaseRenewSecret,
cancel_secret=LeaseCancelSecret,
sharenums=SetOf(int, maxLength=MAX_BUCKETS),
allocated_size=Offset, canary=Referenceable):
"""
@param storage_index: the index of the bucket to be created or
increfed.
@param sharenums: these are the share numbers (probably between 0 and
99) that the sender is proposing to store on this
server.
@param renew_secret: This is the secret used to protect bucket refresh
This secret is generated by the client and
stored for later comparison by the server. Each
server is given a different secret.
@param cancel_secret: This no longer allows lease cancellation, but
must still be a unique value identifying the
lease. XXX stop relying on it to be unique.
@param canary: If the canary is lost before close(), the bucket is
deleted.
@return: tuple of (alreadygot, allocated), where alreadygot is what we
already have and allocated is what we hereby agree to accept.
New leases are added for shares in both lists.
"""
return TupleOf(SetOf(int, maxLength=MAX_BUCKETS),
DictOf(int, RIBucketWriter, maxKeys=MAX_BUCKETS))
def add_lease(storage_index=StorageIndex,
renew_secret=LeaseRenewSecret,
cancel_secret=LeaseCancelSecret):
"""
Add a new lease on the given bucket. If the renew_secret matches an
existing lease, that lease will be renewed instead. If there is no
bucket for the given storage_index, return silently. (note that in
tahoe-1.3.0 and earlier, IndexError was raised if there was no
bucket)
"""
return Any() # returns None now, but future versions might change
def renew_lease(storage_index=StorageIndex, renew_secret=LeaseRenewSecret):
"""
Renew the lease on a given bucket, resetting the timer to 31 days.
Some networks will use this, some will not. If there is no bucket for
the given storage_index, IndexError will be raised.
For mutable shares, if the given renew_secret does not match an
existing lease, IndexError will be raised with a note listing the
server-nodeids on the existing leases, so leases on migrated shares
can be renewed. For immutable shares, IndexError (without the note)
will be raised.
"""
return Any()
def get_buckets(storage_index=StorageIndex):
return DictOf(int, RIBucketReader, maxKeys=MAX_BUCKETS)
def slot_readv(storage_index=StorageIndex,
shares=ListOf(int), readv=ReadVector):
"""Read a vector from the numbered shares associated with the given
storage index. An empty shares list means to return data from all
known shares. Returns a dictionary with one key per share."""
return DictOf(int, ReadData) # shnum -> results
def slot_testv_and_readv_and_writev(storage_index=StorageIndex,
secrets=TupleOf(WriteEnablerSecret,
LeaseRenewSecret,
LeaseCancelSecret),
tw_vectors=TestAndWriteVectorsForShares,
r_vector=ReadVector,
):
"""
General-purpose test-read-and-set operation for mutable slots:
(1) For submitted shnums, compare the test vectors against extant
shares, or against an empty share for shnums that do not exist.
(2) Use the read vectors to extract "old data" from extant shares.
(3) If all tests in (1) passed, then apply the write vectors
(possibly creating new shares).
(4) Return whether the tests passed, and the "old data", which does
not include any modifications made by the writes.
The operation does not interleave with other operations on the same
shareset.
This method is, um, large. The goal is to allow clients to update all
the shares associated with a mutable file in a single round trip.
@param storage_index: the index of the bucket to be created or
increfed.
@param write_enabler: a secret that is stored along with the slot.
Writes are accepted from any caller who can
present the matching secret. A different secret
should be used for each slot*server pair.
@param renew_secret: This is the secret used to protect bucket refresh
This secret is generated by the client and
stored for later comparison by the server. Each
server is given a different secret.
@param cancel_secret: This no longer allows lease cancellation, but
must still be a unique value identifying the
lease. XXX stop relying on it to be unique.
The 'secrets' argument is a tuple of (write_enabler, renew_secret,
cancel_secret). The first is required to perform any write. The
latter two are used when allocating new shares. To simply acquire a
new lease on existing shares, use an empty testv and an empty writev.
Each share can have a separate test vector (i.e. a list of
comparisons to perform). If all vectors for all shares pass, then all
writes for all shares are recorded. Each comparison is a 4-tuple of
(offset, length, operator, specimen), which effectively does a
bool( (read(offset, length)) OPERATOR specimen ) and only performs
the write if all these evaluate to True. Basic test-and-set uses 'eq'.
Write-if-newer uses a seqnum and (offset, length, 'lt', specimen).
Write-if-same-or-newer uses 'le'.
Reads from the end of the container are truncated, and missing shares
behave like empty ones, so to assert that a share doesn't exist (for
use when creating a new share), use (0, 1, 'eq', '').
The write vector will be applied to the given share, expanding it if
necessary. A write vector applied to a share number that did not
exist previously will cause that share to be created. Write vectors
must not overlap (if they do, this will either cause an error or
apply them in an unspecified order). Duplicate write vectors, with
the same offset and data, are currently tolerated but are not
desirable.
In Tahoe-LAFS v1.8.3 or later (except 1.9.0a1), if you send a write
vector whose offset is beyond the end of the current data, the space
between the end of the current data and the beginning of the write
vector will be filled with zero bytes. In earlier versions the
contents of this space was unspecified (and might end up containing
secrets). Storage servers with the new zero-filling behavior will
advertise a true value for the 'fills-holes-with-zero-bytes' key
(under 'http://allmydata.org/tahoe/protocols/storage/v1') in their
version information.
Each write vector is accompanied by a 'new_length' argument, which
can be used to truncate the data. If new_length is not None and it is
less than the current size of the data (after applying all write
vectors), then the data will be truncated to new_length. If
new_length==0, the share will be deleted.
In Tahoe-LAFS v1.8.2 and earlier, new_length could also be used to
enlarge the file by sending a number larger than the size of the data
after applying all write vectors. That behavior was not used, and as
of Tahoe-LAFS v1.8.3 it no longer works and the new_length is ignored
in that case.
If a storage client knows that the server supports zero-filling, for
example from the 'fills-holes-with-zero-bytes' key in its version
information, it can extend the file efficiently by writing a single
zero byte just before the new end-of-file. Otherwise it must
explicitly write zeroes to all bytes between the old and new
end-of-file. In any case it should avoid sending new_length larger
than the size of the data after applying all write vectors.
The read vector is used to extract data from all known shares,
*before* any writes have been applied. The same read vector is used
for all shares. This captures the state that was tested by the test
vector, for extant shares.
This method returns two values: a boolean and a dict. The boolean is
True if the write vectors were applied, False if not. The dict is
keyed by share number, and each value contains a list of strings, one
for each element of the read vector.
If the write_enabler is wrong, this will raise BadWriteEnablerError.
To enable share migration (using update_write_enabler), the exception
will have the nodeid used for the old write enabler embedded in it,
in the following string::
The write enabler was recorded by nodeid '%s'.
Note that the nodeid here is encoded using the same base32 encoding
used by Foolscap and allmydata.util.idlib.nodeid_b2a().
"""
return TupleOf(bool, DictOf(int, ReadData))
def advise_corrupt_share(share_type=str, storage_index=StorageIndex,
shnum=int, reason=str):
"""Clients who discover hash failures in shares that they have
downloaded from me will use this method to inform me about the
failures. I will record their concern so that my operator can
manually inspect the shares in question. I return None.
'share_type' is either 'mutable' or 'immutable'. 'storage_index' is a
(binary) storage index string, and 'shnum' is the integer share
number. 'reason' is a human-readable explanation of the problem,
probably including some expected hash values and the computed ones
that did not match. Corruption advisories for mutable shares should
include a hash of the public key (the same value that appears in the
mutable-file verify-cap), since the current share format does not
store that on disk.
"""
class IStorageBucketWriter(Interface):
"""
Objects of this kind live on the client side.
"""
def put_block(segmentnum, data):
"""
@param segmentnum=int
@param data=ShareData: For most segments, this data will be 'blocksize'
bytes in length. The last segment might be shorter.
@return: a Deferred that fires (with None) when the operation completes
"""
def put_crypttext_hashes(hashes):
"""
@param hashes=ListOf(Hash)
@return: a Deferred that fires (with None) when the operation completes
"""
def put_block_hashes(blockhashes):
"""
@param blockhashes=ListOf(Hash)
@return: a Deferred that fires (with None) when the operation completes
"""
def put_share_hashes(sharehashes):
"""
@param sharehashes=ListOf(TupleOf(int, Hash))
@return: a Deferred that fires (with None) when the operation completes
"""
def put_uri_extension(data):
"""This block of data contains integrity-checking information (hashes
of plaintext, crypttext, and shares), as well as encoding parameters
that are necessary to recover the data. This is a serialized dict
mapping strings to other strings. The hash of this data is kept in
the URI and verified before any of the data is used. All buckets for
a given file contain identical copies of this data.
The serialization format is specified with the following pseudocode:
for k in sorted(dict.keys()):
assert re.match(r'^[a-zA-Z_\-]+$', k)
write(k + ':' + netstring(dict[k]))
@param data=URIExtensionData
@return: a Deferred that fires (with None) when the operation completes
"""
def close():
"""Finish writing and close the bucket. The share is not finalized
until this method is called: if the uploading client disconnects
before calling close(), the partially-written share will be
discarded.
@return: a Deferred that fires (with None) when the operation completes
"""
class IStorageBucketReader(Interface):
def get_block_data(blocknum, blocksize, size):
"""Most blocks will be the same size. The last block might be shorter
than the others.
@param blocknum=int
@param blocksize=int
@param size=int
@return: ShareData
"""
def get_crypttext_hashes():
"""
@return: ListOf(Hash)
"""
def get_block_hashes(at_least_these=()):
"""
@param at_least_these=SetOf(int)
@return: ListOf(Hash)
"""
def get_share_hashes():
"""
@return: ListOf(TupleOf(int, Hash))
"""
def get_uri_extension():
"""
@return: URIExtensionData
"""
class IStorageBroker(Interface):
def get_servers_for_psi(peer_selection_index):
"""
@return: list of IServer instances
"""
def get_connected_servers():
"""
@return: frozenset of connected IServer instances
"""
def get_known_servers():
"""
@return: frozenset of IServer instances
"""
def get_all_serverids():
"""
@return: frozenset of serverid strings
"""
def get_nickname_for_serverid(serverid):
"""
@return: unicode nickname, or None
"""
# methods moved from IntroducerClient, need review
def get_all_connections():
"""Return a frozenset of (nodeid, service_name, rref) tuples, one for
each active connection we've established to a remote service. This is
mostly useful for unit tests that need to wait until a certain number
of connections have been made."""
def get_all_connectors():
"""Return a dict that maps from (nodeid, service_name) to a
RemoteServiceConnector instance for all services that we are actively
trying to connect to. Each RemoteServiceConnector has the following
public attributes::
service_name: the type of service provided, like 'storage'
last_connect_time: when we last established a connection
last_loss_time: when we last lost a connection
version: the peer's version, from the most recent connection
oldest_supported: the peer's oldest supported version, same
rref: the RemoteReference, if connected, otherwise None
remote_host: the IAddress, if connected, otherwise None
This method is intended for monitoring interfaces, such as a web page
that describes connecting and connected peers.
"""
def get_all_peerids():
"""Return a frozenset of all peerids to whom we have a connection (to
one or more services) established. Mostly useful for unit tests."""
def get_all_connections_for(service_name):
"""Return a frozenset of (nodeid, service_name, rref) tuples, one
for each active connection that provides the given SERVICE_NAME."""
def get_permuted_peers(service_name, key):
"""Returns an ordered list of (peerid, rref) tuples, selecting from
the connections that provide SERVICE_NAME, using a hash-based
permutation keyed by KEY. This randomizes the service list in a
repeatable way, to distribute load over many peers.
"""
class IDisplayableServer(Interface):
def get_nickname():
pass
def get_name():
pass
def get_longname():
pass
class IServer(IDisplayableServer):
"""I live in the client, and represent a single server."""
def start_connecting(trigger_cb):
pass
def get_rref():
"""Once a server is connected, I return a RemoteReference.
Before a server is connected for the first time, I return None.
Note that the rref I return will start producing DeadReferenceErrors
once the connection is lost.
"""
class IMutableSlotWriter(Interface):
"""
The interface for a writer around a mutable slot on a remote server.
"""
def set_checkstring(seqnum_or_checkstring, root_hash=None, salt=None):
"""
Set the checkstring that I will pass to the remote server when
writing.
@param checkstring A packed checkstring to use.
Note that implementations can differ in which semantics they
wish to support for set_checkstring -- they can, for example,
build the checkstring themselves from its constituents, or
some other thing.
"""
def get_checkstring():
"""
Get the checkstring that I think currently exists on the remote
server.
"""
def put_block(data, segnum, salt):
"""
Add a block and salt to the share.
"""
def put_encprivkey(encprivkey):
"""
Add the encrypted private key to the share.
"""
def put_blockhashes(blockhashes):
"""
@param blockhashes=list
Add the block hash tree to the share.
"""
def put_sharehashes(sharehashes):
"""
@param sharehashes=dict
Add the share hash chain to the share.
"""
def get_signable():
"""
Return the part of the share that needs to be signed.
"""
def put_signature(signature):
"""
Add the signature to the share.
"""
def put_verification_key(verification_key):
"""
Add the verification key to the share.
"""
def finish_publishing():
"""
Do anything necessary to finish writing the share to a remote
server. I require that no further publishing needs to take place
after this method has been called.
"""
class IURI(Interface):
def init_from_string(uri):
"""Accept a string (as created by my to_string() method) and populate
this instance with its data. I am not normally called directly,
please use the module-level uri.from_string() function to convert
arbitrary URI strings into IURI-providing instances."""
def is_readonly():
"""Return False if this URI be used to modify the data. Return True
if this URI cannot be used to modify the data."""
def is_mutable():
"""Return True if the data can be modified by *somebody* (perhaps
someone who has a more powerful URI than this one)."""
# TODO: rename to get_read_cap()
def get_readonly():
"""Return another IURI instance that represents a read-only form of
this one. If is_readonly() is True, this returns self."""
def get_verify_cap():
"""Return an instance that provides IVerifierURI, which can be used
to check on the availability of the file or directory, without
providing enough capabilities to actually read or modify the
contents. This may return None if the file does not need checking or
verification (e.g. LIT URIs).
"""
def to_string():
"""Return a string of printable ASCII characters, suitable for
passing into init_from_string."""
class IVerifierURI(Interface, IURI):
def init_from_string(uri):
"""Accept a string (as created by my to_string() method) and populate
this instance with its data. I am not normally called directly,
please use the module-level uri.from_string() function to convert
arbitrary URI strings into IURI-providing instances."""
def to_string():
"""Return a string of printable ASCII characters, suitable for
passing into init_from_string."""
class IDirnodeURI(Interface):
"""I am a URI that represents a dirnode."""
class IFileURI(Interface):
"""I am a URI that represents a filenode."""
def get_size():
"""Return the length (in bytes) of the file that I represent."""
class IImmutableFileURI(IFileURI):
pass
class IMutableFileURI(Interface):
pass
class IDirectoryURI(Interface):
pass
class IReadonlyDirectoryURI(Interface):
pass
class CapConstraintError(Exception):
"""A constraint on a cap was violated."""
class MustBeDeepImmutableError(CapConstraintError):
"""Mutable children cannot be added to an immutable directory.
Also, caps obtained from an immutable directory can trigger this error
if they are later found to refer to a mutable object and then used."""
class MustBeReadonlyError(CapConstraintError):
"""Known write caps cannot be specified in a ro_uri field. Also,
caps obtained from a ro_uri field can trigger this error if they
are later found to be write caps and then used."""
class MustNotBeUnknownRWError(CapConstraintError):
"""Cannot add an unknown child cap specified in a rw_uri field."""
class IProgress(Interface):
"""
Remembers progress measured in arbitrary units. Users of these
instances must call ``set_progress_total`` at least once before
progress can be valid, and must use the same units for both
``set_progress_total`` and ``set_progress calls``.
See also:
:class:`allmydata.util.progress.PercentProgress`
"""
progress = Attribute(
"Current amount of progress (in percentage)"
)
def set_progress(self, value):
"""
Sets the current amount of progress.
Arbitrary units, but must match units used for
set_progress_total.
"""
def set_progress_total(self, value):
"""
Sets the total amount of expected progress
Arbitrary units, but must be same units as used when calling
set_progress() on this instance)..
"""
class IReadable(Interface):
"""I represent a readable object -- either an immutable file, or a
specific version of a mutable file.
"""
def is_readonly():
"""Return True if this reference provides mutable access to the given
file or directory (i.e. if you can modify it), or False if not. Note
that even if this reference is read-only, someone else may hold a
read-write reference to it.
For an IReadable returned by get_best_readable_version(), this will
always return True, but for instances of subinterfaces such as
IMutableFileVersion, it may return False."""
def is_mutable():
"""Return True if this file or directory is mutable (by *somebody*,
not necessarily you), False if it is is immutable. Note that a file
might be mutable overall, but your reference to it might be
read-only. On the other hand, all references to an immutable file
will be read-only; there are no read-write references to an immutable
file."""
def get_storage_index():
"""Return the storage index of the file."""
def get_size():
"""Return the length (in bytes) of this readable object."""
def download_to_data(progress=None):
"""Download all of the file contents. I return a Deferred that fires
with the contents as a byte string.
:param progress: None or IProgress implementer
"""
def read(consumer, offset=0, size=None):
"""Download a portion (possibly all) of the file's contents, making
them available to the given IConsumer. Return a Deferred that fires
(with the consumer) when the consumer is unregistered (either because
the last byte has been given to it, or because the consumer threw an
exception during write(), possibly because it no longer wants to
receive data). The portion downloaded will start at 'offset' and
contain 'size' bytes (or the remainder of the file if size==None). It
is an error to read beyond the end of the file: callers must use
get_size() and clip any non-default offset= and size= parameters. It
is permissible to read zero bytes.
The consumer will be used in non-streaming mode: an IPullProducer
will be attached to it.
The consumer will not receive data right away: several network trips
must occur first. The order of events will be::
consumer.registerProducer(p, streaming)
(if streaming == False)::
consumer does p.resumeProducing()
consumer.write(data)
consumer does p.resumeProducing()
consumer.write(data).. (repeat until all data is written)
consumer.unregisterProducer()
deferred.callback(consumer)
If a download error occurs, or an exception is raised by
consumer.registerProducer() or consumer.write(), I will call
consumer.unregisterProducer() and then deliver the exception via
deferred.errback(). To cancel the download, the consumer should call
p.stopProducing(), which will result in an exception being delivered
via deferred.errback().
See src/allmydata/util/consumer.py for an example of a simple
download-to-memory consumer.
"""
class IWriteable(Interface):
"""
I define methods that callers can use to update SDMF and MDMF
mutable files on a Tahoe-LAFS grid.
"""
# XXX: For the moment, we have only this. It is possible that we
# want to move overwrite() and modify() in here too.
def update(data, offset):
"""
I write the data from my data argument to the MDMF file,
starting at offset. I continue writing data until my data
argument is exhausted, appending data to the file as necessary.
"""
# assert IMutableUploadable.providedBy(data)
# to append data: offset=node.get_size_of_best_version()
# do we want to support compacting MDMF?
# for an MDMF file, this can be done with O(data.get_size())
# memory. For an SDMF file, any modification takes
# O(node.get_size_of_best_version()).
class IMutableFileVersion(IReadable):
"""I provide access to a particular version of a mutable file. The
access is read/write if I was obtained from a filenode derived from
a write cap, or read-only if the filenode was derived from a read cap.
"""
def get_sequence_number():
"""Return the sequence number of this version."""
def get_servermap():
"""Return the IMutableFileServerMap instance that was used to create
this object.
"""
def get_writekey():
"""Return this filenode's writekey, or None if the node does not have
write-capability. This may be used to assist with data structures
that need to make certain data available only to writers, such as the
read-write child caps in dirnodes. The recommended process is to have
reader-visible data be submitted to the filenode in the clear (where
it will be encrypted by the filenode using the readkey), but encrypt
writer-visible data using this writekey.
"""
def overwrite(new_contents):
"""Replace the contents of the mutable file, provided that no other
node has published (or is attempting to publish, concurrently) a
newer version of the file than this one.
I will avoid modifying any share that is different than the version
given by get_sequence_number(). However, if another node is writing
to the file at the same time as me, I may manage to update some shares
while they update others. If I see any evidence of this, I will signal
UncoordinatedWriteError, and the file will be left in an inconsistent
state (possibly the version you provided, possibly the old version,
possibly somebody else's version, and possibly a mix of shares from
all of these).
The recommended response to UncoordinatedWriteError is to either
return it to the caller (since they failed to coordinate their
writes), or to attempt some sort of recovery. It may be sufficient to
wait a random interval (with exponential backoff) and repeat your
operation. If I do not signal UncoordinatedWriteError, then I was
able to write the new version without incident.
I return a Deferred that fires (with a PublishStatus object) when the
update has completed.
"""
def modify(modifier_cb):
"""Modify the contents of the file, by downloading this version,
applying the modifier function (or bound method), then uploading
the new version. This will succeed as long as no other node
publishes a version between the download and the upload.
I return a Deferred that fires (with a PublishStatus object) when
the update is complete.
The modifier callable will be given three arguments: a string (with
the old contents), a 'first_time' boolean, and a servermap. As with
download_to_data(), the old contents will be from this version,
but the modifier can use the servermap to make other decisions
(such as refusing to apply the delta if there are multiple parallel
versions, or if there is evidence of a newer unrecoverable version).
'first_time' will be True the first time the modifier is called,
and False on any subsequent calls.
The callable should return a string with the new contents. The
callable must be prepared to be called multiple times, and must
examine the input string to see if the change that it wants to make
is already present in the old version. If it does not need to make
any changes, it can either return None, or return its input string.
If the modifier raises an exception, it will be returned in the
errback.
"""
# The hierarchy looks like this:
# IFilesystemNode
# IFileNode
# IMutableFileNode
# IImmutableFileNode
# IDirectoryNode
class IFilesystemNode(Interface):
def get_cap():
"""Return the strongest 'cap instance' associated with this node.
(writecap for writeable-mutable files/directories, readcap for
immutable or readonly-mutable files/directories). To convert this
into a string, call .to_string() on the result."""
def get_readcap():
"""Return a readonly cap instance for this node. For immutable or
readonly nodes, get_cap() and get_readcap() return the same thing."""
def get_repair_cap():
"""Return an IURI instance that can be used to repair the file, or
None if this node cannot be repaired (either because it is not
distributed, like a LIT file, or because the node does not represent
sufficient authority to create a repair-cap, like a read-only RSA
mutable file node [which cannot create the correct write-enablers]).
"""
def get_verify_cap():
"""Return an IVerifierURI instance that represents the
'verifiy/refresh capability' for this node. The holder of this
capability will be able to renew the lease for this node, protecting
it from garbage-collection. They will also be able to ask a server if
it holds a share for the file or directory.
"""
def get_uri():
"""Return the URI string corresponding to the strongest cap associated
with this node. If this node is read-only, the URI will only offer
read-only access. If this node is read-write, the URI will offer
read-write access.
If you have read-write access to a node and wish to share merely
read-only access with others, use get_readonly_uri().
"""
def get_write_uri():
"""Return the URI string that can be used by others to get write
access to this node, if it is writeable. If this is a read-only node,
return None."""
def get_readonly_uri():
"""Return the URI string that can be used by others to get read-only
access to this node. The result is a read-only URI, regardless of
whether this node is read-only or read-write.
If you have merely read-only access to this node, get_readonly_uri()
will return the same thing as get_uri().
"""
def get_storage_index():
"""Return a string with the (binary) storage index in use on this
download. This may be None if there is no storage index (i.e. LIT
files and directories)."""
def is_readonly():
"""Return True if this reference provides mutable access to the given
file or directory (i.e. if you can modify it), or False if not. Note
that even if this reference is read-only, someone else may hold a
read-write reference to it."""
def is_mutable():
"""Return True if this file or directory is mutable (by *somebody*,
not necessarily you), False if it is is immutable. Note that a file
might be mutable overall, but your reference to it might be
read-only. On the other hand, all references to an immutable file
will be read-only; there are no read-write references to an immutable
file.
"""
def is_unknown():
"""Return True if this is an unknown node."""
def is_allowed_in_immutable_directory():
"""Return True if this node is allowed as a child of a deep-immutable
directory. This is true if either the node is of a known-immutable type,
or it is unknown and read-only.
"""
def raise_error():
"""Raise any error associated with this node."""
# XXX: These may not be appropriate outside the context of an IReadable.
def get_size():
"""Return the length (in bytes) of the data this node represents. For
directory nodes, I return the size of the backing store. I return
synchronously and do not consult the network, so for mutable objects,
I will return the most recently observed size for the object, or None
if I don't remember a size. Use get_current_size, which returns a
Deferred, if you want more up-to-date information."""
def get_current_size():
"""I return a Deferred that fires with the length (in bytes) of the
data this node represents.
"""
class IFileNode(IFilesystemNode):
"""I am a node that represents a file: a sequence of bytes. I am not a
container, like IDirectoryNode."""
def get_best_readable_version():
"""Return a Deferred that fires with an IReadable for the 'best'
available version of the file. The IReadable provides only read
access, even if this filenode was derived from a write cap.
For an immutable file, there is only one version. For a mutable
file, the 'best' version is the recoverable version with the
highest sequence number. If no uncoordinated writes have occurred,
and if enough shares are available, then this will be the most
recent version that has been uploaded. If no version is recoverable,
the Deferred will errback with an UnrecoverableFileError.
"""
def download_best_version(progress=None):
"""Download the contents of the version that would be returned
by get_best_readable_version(). This is equivalent to calling
download_to_data() on the IReadable given by that method.
progress is anything that implements IProgress
I return a Deferred that fires with a byte string when the file
has been fully downloaded. To support streaming download, use
the 'read' method of IReadable. If no version is recoverable,
the Deferred will errback with an UnrecoverableFileError.
"""
def get_size_of_best_version():
"""Find the size of the version that would be returned by
get_best_readable_version().
I return a Deferred that fires with an integer. If no version
is recoverable, the Deferred will errback with an
UnrecoverableFileError.
"""
class IImmutableFileNode(IFileNode, IReadable):
"""I am a node representing an immutable file. Immutable files have
only one version"""
class IMutableFileNode(IFileNode):
"""I provide access to a 'mutable file', which retains its identity
regardless of what contents are put in it.
The consistency-vs-availability problem means that there might be
multiple versions of a file present in the grid, some of which might be
unrecoverable (i.e. have fewer than 'k' shares). These versions are
loosely ordered: each has a sequence number and a hash, and any version
with seqnum=N was uploaded by a node that has seen at least one version
with seqnum=N-1.
The 'servermap' (an instance of IMutableFileServerMap) is used to
describe the versions that are known to be present in the grid, and which
servers are hosting their shares. It is used to represent the 'state of
the world', and is used for this purpose by my test-and-set operations.
Downloading the contents of the mutable file will also return a
servermap. Uploading a new version into the mutable file requires a
servermap as input, and the semantics of the replace operation is
'replace the file with my new version if it looks like nobody else has
changed the file since my previous download'. Because the file is
distributed, this is not a perfect test-and-set operation, but it will do
its best. If the replace process sees evidence of a simultaneous write,
it will signal an UncoordinatedWriteError, so that the caller can take
corrective action.
Most readers will want to use the 'best' current version of the file, and
should use my 'download_best_version()' method.
To unconditionally replace the file, callers should use overwrite(). This
is the mode that user-visible mutable files will probably use.
To apply some delta to the file, call modify() with a callable modifier
function that can apply the modification that you want to make. This is
the mode that dirnodes will use, since most directory modification
operations can be expressed in terms of deltas to the directory state.
Three methods are available for users who need to perform more complex
operations. The first is get_servermap(), which returns an up-to-date
servermap using a specified mode. The second is download_version(), which
downloads a specific version (not necessarily the 'best' one). The third
is 'upload', which accepts new contents and a servermap (which must have
been updated with MODE_WRITE). The upload method will attempt to apply
the new contents as long as no other node has modified the file since the
servermap was updated. This might be useful to a caller who wants to
merge multiple versions into a single new one.
Note that each time the servermap is updated, a specific 'mode' is used,
which determines how many peers are queried. To use a servermap for my
replace() method, that servermap must have been updated in MODE_WRITE.
These modes are defined in allmydata.mutable.common, and consist of
MODE_READ, MODE_WRITE, MODE_ANYTHING, and MODE_CHECK. Please look in
allmydata/mutable/servermap.py for details about the differences.
Mutable files are currently limited in size (about 3.5MB max) and can
only be retrieved and updated all-at-once, as a single big string. Future
versions of our mutable files will remove this restriction.
"""
def get_best_mutable_version():
"""Return a Deferred that fires with an IMutableFileVersion for
the 'best' available version of the file. The best version is
the recoverable version with the highest sequence number. If no
uncoordinated writes have occurred, and if enough shares are
available, then this will be the most recent version that has
been uploaded.
If no version is recoverable, the Deferred will errback with an
UnrecoverableFileError.
"""
def overwrite(new_contents):
"""Unconditionally replace the contents of the mutable file with new
ones. This simply chains get_servermap(MODE_WRITE) and upload(). This
is only appropriate to use when the new contents of the file are
completely unrelated to the old ones, and you do not care about other
clients' changes.
I return a Deferred that fires (with a PublishStatus object) when the
update has completed.
"""
def modify(modifier_cb):
"""Modify the contents of the file, by downloading the current
version, applying the modifier function (or bound method), then
uploading the new version. I return a Deferred that fires (with a
PublishStatus object) when the update is complete.
The modifier callable will be given three arguments: a string (with
the old contents), a 'first_time' boolean, and a servermap. As with
download_best_version(), the old contents will be from the best
recoverable version, but the modifier can use the servermap to make
other decisions (such as refusing to apply the delta if there are
multiple parallel versions, or if there is evidence of a newer
unrecoverable version). 'first_time' will be True the first time the
modifier is called, and False on any subsequent calls.
The callable should return a string with the new contents. The
callable must be prepared to be called multiple times, and must
examine the input string to see if the change that it wants to make
is already present in the old version. If it does not need to make
any changes, it can either return None, or return its input string.
If the modifier raises an exception, it will be returned in the
errback.
"""
def get_servermap(mode):
"""Return a Deferred that fires with an IMutableFileServerMap
instance, updated using the given mode.
"""
def download_version(servermap, version):
"""Download a specific version of the file, using the servermap
as a guide to where the shares are located.
I return a Deferred that fires with the requested contents, or
errbacks with UnrecoverableFileError. Note that a servermap that was
updated with MODE_ANYTHING or MODE_READ may not know about shares for
all versions (those modes stop querying servers as soon as they can
fulfil their goals), so you may want to use MODE_CHECK (which checks
everything) to get increased visibility.
"""
def upload(new_contents, servermap, progress=None):
"""Replace the contents of the file with new ones. This requires a
servermap that was previously updated with MODE_WRITE.
I attempt to provide test-and-set semantics, in that I will avoid
modifying any share that is different than the version I saw in the
servermap. However, if another node is writing to the file at the
same time as me, I may manage to update some shares while they update
others. If I see any evidence of this, I will signal
UncoordinatedWriteError, and the file will be left in an inconsistent
state (possibly the version you provided, possibly the old version,
possibly somebody else's version, and possibly a mix of shares from
all of these).
The recommended response to UncoordinatedWriteError is to either
return it to the caller (since they failed to coordinate their
writes), or to attempt some sort of recovery. It may be sufficient to
wait a random interval (with exponential backoff) and repeat your
operation. If I do not signal UncoordinatedWriteError, then I was
able to write the new version without incident.
``progress`` is either None or an IProgress provider
I return a Deferred that fires (with a PublishStatus object) when the
publish has completed. I will update the servermap in-place with the
location of all new shares.
"""
def get_writekey():
"""Return this filenode's writekey, or None if the node does not have
write-capability. This may be used to assist with data structures
that need to make certain data available only to writers, such as the
read-write child caps in dirnodes. The recommended process is to have
reader-visible data be submitted to the filenode in the clear (where
it will be encrypted by the filenode using the readkey), but encrypt
writer-visible data using this writekey.
"""
def get_version():
"""Returns the mutable file protocol version."""
class NotEnoughSharesError(Exception):
"""Download was unable to get enough shares"""
class NoSharesError(Exception):
"""Download was unable to get any shares at all."""
class DownloadStopped(Exception):
pass
class UploadUnhappinessError(Exception):
"""Upload was unable to satisfy 'servers_of_happiness'"""
class UnableToFetchCriticalDownloadDataError(Exception):
"""I was unable to fetch some piece of critical data that is supposed to
be identically present in all shares."""
class NoServersError(Exception):
"""Upload wasn't given any servers to work with, usually indicating a
network or Introducer problem."""
class ExistingChildError(Exception):
"""A directory node was asked to add or replace a child that already
exists, and overwrite= was set to False."""
class NoSuchChildError(Exception):
"""A directory node was asked to fetch a child that does not exist."""
def __str__(self):
# avoid UnicodeEncodeErrors when converting to str
return self.__repr__()
class ChildOfWrongTypeError(Exception):
"""An operation was attempted on a child of the wrong type (file or directory)."""
class IDirectoryNode(IFilesystemNode):
"""I represent a filesystem node that is a container, with a
name-to-child mapping, holding the tahoe equivalent of a directory. All
child names are unicode strings, and all children are some sort of
IFilesystemNode (a file, subdirectory, or unknown node).
"""
def get_uri():
"""
The dirnode ('1') URI returned by this method can be used in
set_uri() on a different directory ('2') to 'mount' a reference to
this directory ('1') under the other ('2'). This URI is just a
string, so it can be passed around through email or other out-of-band
protocol.
"""
def get_readonly_uri():
"""
The dirnode ('1') URI returned by this method can be used in
set_uri() on a different directory ('2') to 'mount' a reference to
this directory ('1') under the other ('2'). This URI is just a
string, so it can be passed around through email or other out-of-band
protocol.
"""
def list():
"""I return a Deferred that fires with a dictionary mapping child
name (a unicode string) to (node, metadata_dict) tuples, in which
'node' is an IFilesystemNode and 'metadata_dict' is a dictionary of
metadata."""
def has_child(name):
"""I return a Deferred that fires with a boolean, True if there
exists a child of the given name, False if not. The child name must
be a unicode string."""
def get(name):
"""I return a Deferred that fires with a specific named child node,
which is an IFilesystemNode. The child name must be a unicode string.
I raise NoSuchChildError if I do not have a child by that name."""
def get_metadata_for(name):
"""I return a Deferred that fires with the metadata dictionary for
a specific named child node. The child name must be a unicode string.
This metadata is stored in the *edge*, not in the child, so it is
attached to the parent dirnode rather than the child node.
I raise NoSuchChildError if I do not have a child by that name."""
def set_metadata_for(name, metadata):
"""I replace any existing metadata for the named child with the new
metadata. The child name must be a unicode string. This metadata is
stored in the *edge*, not in the child, so it is attached to the
parent dirnode rather than the child node. I return a Deferred
(that fires with this dirnode) when the operation is complete.
I raise NoSuchChildError if I do not have a child by that name."""
def get_child_at_path(path):
"""Transform a child path into an IFilesystemNode.
I perform a recursive series of 'get' operations to find the named
descendant node. I return a Deferred that fires with the node, or
errbacks with NoSuchChildError if the node could not be found.
The path can be either a single string (slash-separated) or a list of
path-name elements. All elements must be unicode strings.
"""
def get_child_and_metadata_at_path(path):
"""Transform a child path into an IFilesystemNode and metadata.
I am like get_child_at_path(), but my Deferred fires with a tuple of
(node, metadata). The metadata comes from the last edge. If the path
is empty, the metadata will be an empty dictionary.
"""
def set_uri(name, writecap, readcap=None, metadata=None, overwrite=True):
"""I add a child (by writecap+readcap) at the specific name. I return
a Deferred that fires when the operation finishes. If overwrite= is
True, I will replace any existing child of the same name, otherwise
an existing child will cause me to return ExistingChildError. The
child name must be a unicode string.
The child caps could be for a file, or for a directory. If you have
both the writecap and readcap, you should provide both arguments.
If you have only one cap and don't know whether it is read-only,
provide it as the writecap argument and leave the readcap as None.
If you have only one cap that is known to be read-only, provide it
as the readcap argument and leave the writecap as None.
The filecaps are typically obtained from an IFilesystemNode with
get_uri() and get_readonly_uri().
If metadata= is provided, I will use it as the metadata for the named
edge. This will replace any existing metadata. If metadata= is left
as the default value of None, I will set ['mtime'] to the current
time, and I will set ['ctime'] to the current time if there was not
already a child by this name present. This roughly matches the
ctime/mtime semantics of traditional filesystems. See the
"About the metadata" section of webapi.txt for futher information.
If this directory node is read-only, the Deferred will errback with a
NotWriteableError."""
def set_children(entries, overwrite=True):
"""Add multiple children (by writecap+readcap) to a directory node.
Takes a dictionary, with childname as keys and (writecap, readcap)
tuples (or (writecap, readcap, metadata) triples) as values. Returns
a Deferred that fires (with this dirnode) when the operation
finishes. This is equivalent to calling set_uri() multiple times, but
is much more efficient. All child names must be unicode strings.
"""
def set_node(name, child, metadata=None, overwrite=True):
"""I add a child at the specific name. I return a Deferred that fires
when the operation finishes. This Deferred will fire with the child
node that was just added. I will replace any existing child of the
same name. The child name must be a unicode string. The 'child'
instance must be an instance providing IFilesystemNode.
If metadata= is provided, I will use it as the metadata for the named
edge. This will replace any existing metadata. If metadata= is left
as the default value of None, I will set ['mtime'] to the current
time, and I will set ['ctime'] to the current time if there was not
already a child by this name present. This roughly matches the
ctime/mtime semantics of traditional filesystems. See the
"About the metadata" section of webapi.txt for futher information.
If this directory node is read-only, the Deferred will errback with a
NotWriteableError."""
def set_nodes(entries, overwrite=True):
"""Add multiple children to a directory node. Takes a dict mapping
unicode childname to (child_node, metdata) tuples. If metdata=None,
the original metadata is left unmodified. Returns a Deferred that
fires (with this dirnode) when the operation finishes. This is
equivalent to calling set_node() multiple times, but is much more
efficient."""
def add_file(name, uploadable, metadata=None, overwrite=True, progress=None):
"""I upload a file (using the given IUploadable), then attach the
resulting ImmutableFileNode to the directory at the given name. I set
metadata the same way as set_uri and set_node. The child name must be
a unicode string.
``progress`` either provides IProgress or is None
I return a Deferred that fires (with the IFileNode of the uploaded
file) when the operation completes."""
def delete(name, must_exist=True, must_be_directory=False, must_be_file=False):
"""I remove the child at the specific name. I return a Deferred that
fires when the operation finishes. The child name must be a unicode
string. If must_exist is True and I do not have a child by that name,
I raise NoSuchChildError. If must_be_directory is True and the child
is a file, or if must_be_file is True and the child is a directory,
I raise ChildOfWrongTypeError."""
def create_subdirectory(name, initial_children={}, overwrite=True,
mutable=True, mutable_version=None, metadata=None):
"""I create and attach a directory at the given name. The new
directory can be empty, or it can be populated with children
according to 'initial_children', which takes a dictionary in the same
format as set_nodes (i.e. mapping unicode child name to (childnode,
metadata) tuples). The child name must be a unicode string. I return
a Deferred that fires (with the new directory node) when the
operation finishes."""
def move_child_to(current_child_name, new_parent, new_child_name=None,
overwrite=True):
"""I take one of my children and move them to a new parent. The child
is referenced by name. On the new parent, the child will live under
'new_child_name', which defaults to 'current_child_name'. TODO: what
should we do about metadata? I return a Deferred that fires when the
operation finishes. The child name must be a unicode string. I raise
NoSuchChildError if I do not have a child by that name."""
def build_manifest():
"""I generate a table of everything reachable from this directory.
I also compute deep-stats as described below.
I return a Monitor. The Monitor's results will be a dictionary with
four elements:
res['manifest']: a list of (path, cap) tuples for all nodes
(directories and files) reachable from this one.
'path' will be a tuple of unicode strings. The
origin dirnode will be represented by an empty path
tuple.
res['verifycaps']: a list of (printable) verifycap strings, one for
each reachable non-LIT node. This is a set:
it will contain no duplicates.
res['storage-index']: a list of (base32) storage index strings,
one for each reachable non-LIT node. This is
a set: it will contain no duplicates.
res['stats']: a dictionary, the same that is generated by
start_deep_stats() below.
The Monitor will also have an .origin_si attribute with the (binary)
storage index of the starting point.
"""
def start_deep_stats():
"""Return a Monitor, examining all nodes (directories and files)
reachable from this one. The Monitor's results will be a dictionary
with the following keys::
count-immutable-files: count of how many CHK files are in the set
count-mutable-files: same, for mutable files (does not include
directories)
count-literal-files: same, for LIT files
count-files: sum of the above three
count-directories: count of directories
size-immutable-files: total bytes for all CHK files in the set
size-mutable-files (TODO): same, for current version of all mutable
files, does not include directories
size-literal-files: same, for LIT files
size-directories: size of mutable files used by directories
largest-directory: number of bytes in the largest directory
largest-directory-children: number of children in the largest
directory
largest-immutable-file: number of bytes in the largest CHK file
size-mutable-files is not yet implemented, because it would involve
even more queries than deep_stats does.
The Monitor will also have an .origin_si attribute with the (binary)
storage index of the starting point.
This operation will visit every directory node underneath this one,
and can take a long time to run. On a typical workstation with good
bandwidth, this can examine roughly 15 directories per second (and
takes several minutes of 100% CPU for ~1700 directories).
"""
class ICodecEncoder(Interface):
def set_params(data_size, required_shares, max_shares):
"""Set up the parameters of this encoder.
This prepares the encoder to perform an operation that converts a
single block of data into a number of shares, such that a future
ICodecDecoder can use a subset of these shares to recover the
original data. This operation is invoked by calling encode(). Once
the encoding parameters are set up, the encode operation can be
invoked multiple times.
set_params() prepares the encoder to accept blocks of input data that
are exactly 'data_size' bytes in length. The encoder will be prepared
to produce 'max_shares' shares for each encode() operation (although
see the 'desired_share_ids' to use less CPU). The encoding math will
be chosen such that the decoder can get by with as few as
'required_shares' of these shares and still reproduce the original
data. For example, set_params(1000, 5, 5) offers no redundancy at
all, whereas set_params(1000, 1, 10) provides 10x redundancy.
Numerical Restrictions: 'data_size' is required to be an integral
multiple of 'required_shares'. In general, the caller should choose
required_shares and max_shares based upon their reliability
requirements and the number of peers available (the total storage
space used is roughly equal to max_shares*data_size/required_shares),
then choose data_size to achieve the memory footprint desired (larger
data_size means more efficient operation, smaller data_size means
smaller memory footprint).
In addition, 'max_shares' must be equal to or greater than
'required_shares'. Of course, setting them to be equal causes
encode() to degenerate into a particularly slow form of the 'split'
utility.
See encode() for more details about how these parameters are used.
set_params() must be called before any other ICodecEncoder methods
may be invoked.
"""
def get_params():
"""Return the 3-tuple of data_size, required_shares, max_shares"""
def get_encoder_type():
"""Return a short string that describes the type of this encoder.
There is required to be a global table of encoder classes. This method
returns an index into this table; the value at this index is an
encoder class, and this encoder is an instance of that class.
"""
def get_block_size():
"""Return the length of the shares that encode() will produce.
"""
def encode_proposal(data, desired_share_ids=None):
"""Encode some data.
'data' must be a string (or other buffer object), and len(data) must
be equal to the 'data_size' value passed earlier to set_params().
This will return a Deferred that will fire with two lists. The first
is a list of shares, each of which is a string (or other buffer
object) such that len(share) is the same as what get_share_size()
returned earlier. The second is a list of shareids, in which each is
an integer. The lengths of the two lists will always be equal to each
other. The user should take care to keep each share closely
associated with its shareid, as one is useless without the other.
The length of this output list will normally be the same as the value
provided to the 'max_shares' parameter of set_params(). This may be
different if 'desired_share_ids' is provided.
'desired_share_ids', if provided, is required to be a sequence of
ints, each of which is required to be >= 0 and < max_shares. If not
provided, encode() will produce 'max_shares' shares, as if
'desired_share_ids' were set to range(max_shares). You might use this
if you initially thought you were going to use 10 peers, started
encoding, and then two of the peers dropped out: you could use
desired_share_ids= to skip the work (both memory and CPU) of
producing shares for the peers that are no longer available.
"""
def encode(inshares, desired_share_ids=None):
"""Encode some data. This may be called multiple times. Each call is
independent.
inshares is a sequence of length required_shares, containing buffers
(i.e. strings), where each buffer contains the next contiguous
non-overlapping segment of the input data. Each buffer is required to
be the same length, and the sum of the lengths of the buffers is
required to be exactly the data_size promised by set_params(). (This
implies that the data has to be padded before being passed to
encode(), unless of course it already happens to be an even multiple
of required_shares in length.)
Note: the requirement to break up your data into
'required_shares' chunks of exactly the right length before
calling encode() is surprising from point of view of a user
who doesn't know how FEC works. It feels like an
implementation detail that has leaked outside the abstraction
barrier. Is there a use case in which the data to be encoded
might already be available in pre-segmented chunks, such that
it is faster or less work to make encode() take a list rather
than splitting a single string?
Yes, there is: suppose you are uploading a file with K=64,
N=128, segsize=262,144. Then each in-share will be of size
4096. If you use this .encode() API then your code could first
read each successive 4096-byte chunk from the file and store
each one in a Python string and store each such Python string
in a Python list. Then you could call .encode(), passing that
list as "inshares". The encoder would generate the other 64
"secondary shares" and return to you a new list containing
references to the same 64 Python strings that you passed in
(as the primary shares) plus references to the new 64 Python
strings.
(You could even imagine that your code could use readv() so
that the operating system can arrange to get all of those
bytes copied from the file into the Python list of Python
strings as efficiently as possible instead of having a loop
written in C or in Python to copy the next part of the file
into the next string.)
On the other hand if you instead use the .encode_proposal()
API (above), then your code can first read in all of the
262,144 bytes of the segment from the file into a Python
string, then call .encode_proposal() passing the segment data
as the "data" argument. The encoder would basically first
split the "data" argument into a list of 64 in-shares of 4096
byte each, and then do the same thing that .encode() does. So
this would result in a little bit more copying of data and a
little bit higher of a "maximum memory usage" during the
process, although it might or might not make a practical
difference for our current use cases.
Note that "inshares" is a strange name for the parameter if
you think of the parameter as being just for feeding in data
to the codec. It makes more sense if you think of the result
of this encoding as being the set of shares from inshares plus
an extra set of "secondary shares" (or "check shares"). It is
a surprising name! If the API is going to be surprising then
the name should be surprising. If we switch to
encode_proposal() above then we should also switch to an
unsurprising name.
'desired_share_ids', if provided, is required to be a sequence of
ints, each of which is required to be >= 0 and < max_shares. If not
provided, encode() will produce 'max_shares' shares, as if
'desired_share_ids' were set to range(max_shares). You might use this
if you initially thought you were going to use 10 peers, started
encoding, and then two of the peers dropped out: you could use
desired_share_ids= to skip the work (both memory and CPU) of
producing shares for the peers that are no longer available.
For each call, encode() will return a Deferred that fires with two
lists, one containing shares and the other containing the shareids.
The get_share_size() method can be used to determine the length of
the share strings returned by encode(). Each shareid is a small
integer, exactly as passed into 'desired_share_ids' (or
range(max_shares), if desired_share_ids was not provided).
The shares and their corresponding shareids are required to be kept
together during storage and retrieval. Specifically, the share data is
useless by itself: the decoder needs to be told which share is which
by providing it with both the shareid and the actual share data.
This function will allocate an amount of memory roughly equal to::
(max_shares - required_shares) * get_share_size()
When combined with the memory that the caller must allocate to
provide the input data, this leads to a memory footprint roughly
equal to the size of the resulting encoded shares (i.e. the expansion
factor times the size of the input segment).
"""
# rejected ideas:
#
# returning a list of (shareidN,shareN) tuples instead of a pair of
# lists (shareids..,shares..). Brian thought the tuples would
# encourage users to keep the share and shareid together throughout
# later processing, Zooko pointed out that the code to iterate
# through two lists is not really more complicated than using a list
# of tuples and there's also a performance improvement
#
# having 'data_size' not required to be an integral multiple of
# 'required_shares'. Doing this would require encode() to perform
# padding internally, and we'd prefer to have any padding be done
# explicitly by the caller. Yes, it is an abstraction leak, but
# hopefully not an onerous one.
class ICodecDecoder(Interface):
def set_params(data_size, required_shares, max_shares):
"""Set the params. They have to be exactly the same ones that were
used for encoding."""
def get_needed_shares():
"""Return the number of shares needed to reconstruct the data.
set_params() is required to be called before this."""
def decode(some_shares, their_shareids):
"""Decode a partial list of shares into data.
'some_shares' is required to be a sequence of buffers of sharedata, a
subset of the shares returned by ICodecEncode.encode(). Each share is
required to be of the same length. The i'th element of their_shareids
is required to be the shareid of the i'th buffer in some_shares.
This returns a Deferred that fires with a sequence of buffers. This
sequence will contain all of the segments of the original data, in
order. The sum of the lengths of all of the buffers will be the
'data_size' value passed into the original ICodecEncode.set_params()
call. To get back the single original input block of data, use
''.join(output_buffers), or you may wish to simply write them in
order to an output file.
Note that some of the elements in the result sequence may be
references to the elements of the some_shares input sequence. In
particular, this means that if those share objects are mutable (e.g.
arrays) and if they are changed, then both the input (the
'some_shares' parameter) and the output (the value given when the
deferred is triggered) will change.
The length of 'some_shares' is required to be exactly the value of
'required_shares' passed into the original ICodecEncode.set_params()
call.
"""
class IEncoder(Interface):
"""I take an object that provides IEncryptedUploadable, which provides
encrypted data, and a list of shareholders. I then encode, hash, and
deliver shares to those shareholders. I will compute all the necessary
Merkle hash trees that are necessary to validate the crypttext that
eventually comes back from the shareholders. I provide the URI Extension
Block Hash, and the encoding parameters, both of which must be included
in the URI.
I do not choose shareholders, that is left to the IUploader. I must be
given a dict of RemoteReferences to storage buckets that are ready and
willing to receive data.
"""
def set_size(size):
"""Specify the number of bytes that will be encoded. This must be
peformed before get_serialized_params() can be called.
"""
def set_encrypted_uploadable(u):
"""Provide a source of encrypted upload data. 'u' must implement
IEncryptedUploadable.
When this is called, the IEncryptedUploadable will be queried for its
length and the storage_index that should be used.
This returns a Deferred that fires with this Encoder instance.
This must be performed before start() can be called.
"""
def get_param(name):
"""Return an encoding parameter, by name.
'storage_index': return a string with the (16-byte truncated SHA-256
hash) storage index to which these shares should be
pushed.
'share_counts': return a tuple describing how many shares are used:
(needed_shares, servers_of_happiness, total_shares)
'num_segments': return an int with the number of segments that
will be encoded.
'segment_size': return an int with the size of each segment.
'block_size': return the size of the individual blocks that will
be delivered to a shareholder's put_block() method. By
knowing this, the shareholder will be able to keep all
blocks in a single file and still provide random access
when reading them. # TODO: can we avoid exposing this?
'share_size': an int with the size of the data that will be stored
on each shareholder. This is aggregate amount of data
that will be sent to the shareholder, summed over all
the put_block() calls I will ever make. It is useful to
determine this size before asking potential
shareholders whether they will grant a lease or not,
since their answers will depend upon how much space we
need. TODO: this might also include some amount of
overhead, like the size of all the hashes. We need to
decide whether this is useful or not.
'serialized_params': a string with a concise description of the
codec name and its parameters. This may be passed
into the IUploadable to let it make sure that
the same file encoded with different parameters
will result in different storage indexes.
Once this is called, set_size() and set_params() may not be called.
"""
def set_shareholders(shareholders, servermap):
"""Tell the encoder where to put the encoded shares. 'shareholders'
must be a dictionary that maps share number (an integer ranging from
0 to n-1) to an instance that provides IStorageBucketWriter.
'servermap' is a dictionary that maps share number (as defined above)
to a set of peerids. This must be performed before start() can be
called."""
def start():
"""Begin the encode/upload process. This involves reading encrypted
data from the IEncryptedUploadable, encoding it, uploading the shares
to the shareholders, then sending the hash trees.
set_encrypted_uploadable() and set_shareholders() must be called
before this can be invoked.
This returns a Deferred that fires with a verify cap when the upload
process is complete. The verifycap, plus the encryption key, is
sufficient to construct the read cap.
"""
class IDecoder(Interface):
"""I take a list of shareholders and some setup information, then
download, validate, decode, and decrypt data from them, writing the
results to an output file.
I do not locate the shareholders, that is left to the IDownloader. I must
be given a dict of RemoteReferences to storage buckets that are ready to
send data.
"""
def setup(outfile):
"""I take a file-like object (providing write and close) to which all
the plaintext data will be written.
TODO: producer/consumer . Maybe write() should return a Deferred that
indicates when it will accept more data? But probably having the
IDecoder be a producer is easier to glue to IConsumer pieces.
"""
def set_shareholders(shareholders):
"""I take a dictionary that maps share identifiers (small integers)
to RemoteReferences that provide RIBucketReader. This must be called
before start()."""
def start():
"""I start the download. This process involves retrieving data and
hash chains from the shareholders, using the hashes to validate the
data, decoding the shares into segments, decrypting the segments,
then writing the resulting plaintext to the output file.
I return a Deferred that will fire (with self) when the download is
complete.
"""
class IDownloadTarget(Interface):
# Note that if the IDownloadTarget is also an IConsumer, the downloader
# will register itself as a producer. This allows the target to invoke
# downloader.pauseProducing, resumeProducing, and stopProducing.
def open(size):
"""Called before any calls to write() or close(). If an error
occurs before any data is available, fail() may be called without
a previous call to open().
'size' is the length of the file being downloaded, in bytes."""
def write(data):
"""Output some data to the target."""
def close():
"""Inform the target that there is no more data to be written."""
def fail(why):
"""fail() is called to indicate that the download has failed. 'why'
is a Failure object indicating what went wrong. No further methods
will be invoked on the IDownloadTarget after fail()."""
def register_canceller(cb):
"""The CiphertextDownloader uses this to register a no-argument function
that the target can call to cancel the download. Once this canceller
is invoked, no further calls to write() or close() will be made."""
def finish():
"""When the CiphertextDownloader is done, this finish() function will be
called. Whatever it returns will be returned to the invoker of
Downloader.download.
"""
class IDownloader(Interface):
def download(uri, target):
"""Perform a CHK download, sending the data to the given target.
'target' must provide IDownloadTarget.
Returns a Deferred that fires (with the results of target.finish)
when the download is finished, or errbacks if something went wrong."""
class IEncryptedUploadable(Interface):
def set_upload_status(upload_status):
"""Provide an IUploadStatus object that should be filled with status
information. The IEncryptedUploadable is responsible for setting
key-determination progress ('chk'), size, storage_index, and
ciphertext-fetch progress. It may delegate some of this
responsibility to others, in particular to the IUploadable."""
def get_size():
"""This behaves just like IUploadable.get_size()."""
def get_all_encoding_parameters():
"""Return a Deferred that fires with a tuple of
(k,happy,n,segment_size). The segment_size will be used as-is, and
must match the following constraints: it must be a multiple of k, and
it shouldn't be unreasonably larger than the file size (if
segment_size is larger than filesize, the difference must be stored
as padding).
This usually passes through to the IUploadable method of the same
name.
The encoder strictly obeys the values returned by this method. To
make an upload use non-default encoding parameters, you must arrange
to control the values that this method returns.
"""
def get_storage_index():
"""Return a Deferred that fires with a 16-byte storage index.
"""
def read_encrypted(length, hash_only):
"""This behaves just like IUploadable.read(), but returns crypttext
instead of plaintext. If hash_only is True, then this discards the
data (and returns an empty list); this improves efficiency when
resuming an interrupted upload (where we need to compute the
plaintext hashes, but don't need the redundant encrypted data)."""
def close():
"""Just like IUploadable.close()."""
class IUploadable(Interface):
def set_upload_status(upload_status):
"""Provide an IUploadStatus object that should be filled with status
information. The IUploadable is responsible for setting
key-determination progress ('chk')."""
def set_default_encoding_parameters(params):
"""Set the default encoding parameters, which must be a dict mapping
strings to ints. The meaningful keys are 'k', 'happy', 'n', and
'max_segment_size'. These might have an influence on the final
encoding parameters returned by get_all_encoding_parameters(), if the
Uploadable doesn't have more specific preferences.
This call is optional: if it is not used, the Uploadable will use
some built-in defaults. If used, this method must be called before
any other IUploadable methods to have any effect.
"""
def get_size():
"""Return a Deferred that will fire with the length of the data to be
uploaded, in bytes. This will be called before the data is actually
used, to compute encoding parameters.
"""
def get_all_encoding_parameters():
"""Return a Deferred that fires with a tuple of
(k,happy,n,segment_size). The segment_size will be used as-is, and
must match the following constraints: it must be a multiple of k, and
it shouldn't be unreasonably larger than the file size (if
segment_size is larger than filesize, the difference must be stored
as padding).
The relative values of k and n allow some IUploadables to request
better redundancy than others (in exchange for consuming more space
in the grid).
Larger values of segment_size reduce hash overhead, while smaller
values reduce memory footprint and cause data to be delivered in
smaller pieces (which may provide a smoother and more predictable
download experience).
The encoder strictly obeys the values returned by this method. To
make an upload use non-default encoding parameters, you must arrange
to control the values that this method returns. One way to influence
them may be to call set_encoding_parameters() before calling
get_all_encoding_parameters().
"""
def get_encryption_key():
"""Return a Deferred that fires with a 16-byte AES key. This key will
be used to encrypt the data. The key will also be hashed to derive
the StorageIndex.
Uploadables that want to achieve convergence should hash their file
contents and the serialized_encoding_parameters to form the key
(which of course requires a full pass over the data). Uploadables can
use the upload.ConvergentUploadMixin class to achieve this
automatically.
Uploadables that do not care about convergence (or do not wish to
make multiple passes over the data) can simply return a
strongly-random 16 byte string.
get_encryption_key() may be called multiple times: the IUploadable is
required to return the same value each time.
"""
def read(length):
"""Return a Deferred that fires with a list of strings (perhaps with
only a single element) that, when concatenated together, contain the
next 'length' bytes of data. If EOF is near, this may provide fewer
than 'length' bytes. The total number of bytes provided by read()
before it signals EOF must equal the size provided by get_size().
If the data must be acquired through multiple internal read
operations, returning a list instead of a single string may help to
reduce string copies. However, the length of the concatenated strings
must equal the amount of data requested, unless EOF is encountered.
Long reads, or short reads without EOF, are not allowed. read()
should return the same amount of data as a local disk file read, just
in a different shape and asynchronously.
'length' will typically be equal to (min(get_size(),1MB)/req_shares),
so a 10kB file means length=3kB, 100kB file means length=30kB,
and >=1MB file means length=300kB.
This method provides for a single full pass through the data. Later
use cases may desire multiple passes or access to only parts of the
data (such as a mutable file making small edits-in-place). This API
will be expanded once those use cases are better understood.
"""
def close():
"""The upload is finished, and whatever filehandle was in use may be
closed."""
class IMutableUploadable(Interface):
"""
I represent content that is due to be uploaded to a mutable filecap.
"""
# This is somewhat simpler than the IUploadable interface above
# because mutable files do not need to be concerned with possibly
# generating a CHK, nor with per-file keys. It is a subset of the
# methods in IUploadable, though, so we could just as well implement
# the mutable uploadables as IUploadables that don't happen to use
# those methods (with the understanding that the unused methods will
# never be called on such objects)
def get_size():
"""
Returns a Deferred that fires with the size of the content held
by the uploadable.
"""
def read(length):
"""
Returns a list of strings that, when concatenated, are the next
length bytes of the file, or fewer if there are fewer bytes
between the current location and the end of the file.
"""
def close():
"""
The process that used the Uploadable is finished using it, so
the uploadable may be closed.
"""
class IUploadResults(Interface):
"""I am returned by immutable upload() methods and contain the results of
the upload.
Note that some of my methods return empty values (0 or an empty dict)
when called for non-distributed LIT files."""
def get_file_size():
"""Return the file size, in bytes."""
def get_uri():
"""Return the (string) URI of the object uploaded, a CHK readcap."""
def get_ciphertext_fetched():
"""Return the number of bytes fetched by the helpe for this upload,
or 0 if the helper did not need to fetch any bytes (or if there was
no helper)."""
def get_preexisting_shares():
"""Return the number of shares that were already present in the grid."""
def get_pushed_shares():
"""Return the number of shares that were uploaded."""
def get_sharemap():
"""Return a dict mapping share identifier to set of IServer
instances. This indicates which servers were given which shares. For
immutable files, the shareid is an integer (the share number, from 0
to N-1). For mutable files, it is a string of the form
'seq%d-%s-sh%d', containing the sequence number, the roothash, and
the share number."""
def get_servermap():
"""Return dict mapping IServer instance to a set of share numbers."""
def get_timings():
"""Return dict of timing information, mapping name to seconds. All
times are floats:
total : total upload time, start to finish
storage_index : time to compute the storage index
peer_selection : time to decide which peers will be used
contacting_helper : initial helper query to upload/no-upload decision
helper_total : initial helper query to helper finished pushing
cumulative_fetch : helper waiting for ciphertext requests
total_fetch : helper start to last ciphertext response
cumulative_encoding : just time spent in zfec
cumulative_sending : just time spent waiting for storage servers
hashes_and_close : last segment push to shareholder close
total_encode_and_push : first encode to shareholder close
"""
def get_uri_extension_data():
"""Return the dict of UEB data created for this file."""
def get_verifycapstr():
"""Return the (string) verify-cap URI for the uploaded object."""
class IDownloadResults(Interface):
"""I am created internally by download() methods. I contain a number of
public attributes that contain details about the download process.::
.file_size : the size of the file, in bytes
.servers_used : set of server peerids that were used during download
.server_problems : dict mapping server peerid to a problem string. Only
servers that had problems (bad hashes, disconnects)
are listed here.
.servermap : dict mapping server peerid to a set of share numbers. Only
servers that had any shares are listed here.
.timings : dict of timing information, mapping name to seconds (float)
peer_selection : time to ask servers about shares
servers_peer_selection : dict of peerid to DYHB-query time
uri_extension : time to fetch a copy of the URI extension block
hashtrees : time to fetch the hash trees
segments : time to fetch, decode, and deliver segments
cumulative_fetch : time spent waiting for storage servers
cumulative_decode : just time spent in zfec
cumulative_decrypt : just time spent in decryption
total : total download time, start to finish
fetch_per_server : dict of server to list of per-segment fetch times
"""
class IUploader(Interface):
def upload(uploadable):
"""Upload the file. 'uploadable' must impement IUploadable. This
returns a Deferred that fires with an IUploadResults instance, from
which the URI of the file can be obtained as results.uri ."""
class ICheckable(Interface):
def check(monitor, verify=False, add_lease=False):
"""Check up on my health, optionally repairing any problems.
This returns a Deferred that fires with an instance that provides
ICheckResults, or None if the object is non-distributed (i.e. LIT
files).
The monitor will be checked periodically to see if the operation has
been cancelled. If so, no new queries will be sent, and the Deferred
will fire (with a OperationCancelledError) immediately.
Filenodes and dirnodes (which provide IFilesystemNode) are also
checkable. Instances that represent verifier-caps will be checkable
but not downloadable. Some objects (like LIT files) do not actually
live in the grid, and their checkers return None (non-distributed
files are always healthy).
If verify=False, a relatively lightweight check will be performed: I
will ask all servers if they have a share for me, and I will believe
whatever they say. If there are at least N distinct shares on the
grid, my results will indicate r.is_healthy()==True. This requires a
roundtrip to each server, but does not transfer very much data, so
the network bandwidth is fairly low.
If verify=True, a more resource-intensive check will be performed:
every share will be downloaded, and the hashes will be validated on
every bit. I will ignore any shares that failed their hash checks. If
there are at least N distinct valid shares on the grid, my results
will indicate r.is_healthy()==True. This requires N/k times as much
download bandwidth (and server disk IO) as a regular download. If a
storage server is holding a corrupt share, or is experiencing memory
failures during retrieval, or is malicious or buggy, then
verification will detect the problem, but checking will not.
If add_lease=True, I will ensure that an up-to-date lease is present
on each share. The lease secrets will be derived from by node secret
(in BASEDIR/private/secret), so either I will add a new lease to the
share, or I will merely renew the lease that I already had. In a
future version of the storage-server protocol (once Accounting has
been implemented), there may be additional options here to define the
kind of lease that is obtained (which account number to claim, etc).
TODO: any problems seen during checking will be reported to the
health-manager.furl, a centralized object that is responsible for
figuring out why files are unhealthy so corrective action can be
taken.
"""
def check_and_repair(monitor, verify=False, add_lease=False):
"""Like check(), but if the file/directory is not healthy, attempt to
repair the damage.
Any non-healthy result will cause an immediate repair operation, to
generate and upload new shares. After repair, the file will be as
healthy as we can make it. Details about what sort of repair is done
will be put in the check-and-repair results. The Deferred will not
fire until the repair is complete.
This returns a Deferred that fires with an instance of
ICheckAndRepairResults."""
class IDeepCheckable(Interface):
def start_deep_check(verify=False, add_lease=False):
"""Check upon the health of me and everything I can reach.
This is a recursive form of check(), useable only on dirnodes.
I return a Monitor, with results that are an IDeepCheckResults
object.
TODO: If any of the directories I traverse are unrecoverable, the
Monitor will report failure. If any of the files I check upon are
unrecoverable, those problems will be reported in the
IDeepCheckResults as usual, and the Monitor will not report a
failure.
"""
def start_deep_check_and_repair(verify=False, add_lease=False):
"""Check upon the health of me and everything I can reach. Repair
anything that isn't healthy.
This is a recursive form of check_and_repair(), useable only on
dirnodes.
I return a Monitor, with results that are an
IDeepCheckAndRepairResults object.
TODO: If any of the directories I traverse are unrecoverable, the
Monitor will report failure. If any of the files I check upon are
unrecoverable, those problems will be reported in the
IDeepCheckResults as usual, and the Monitor will not report a
failure.
"""
class ICheckResults(Interface):
"""I contain the detailed results of a check/verify operation.
"""
def get_storage_index():
"""Return a string with the (binary) storage index."""
def get_storage_index_string():
"""Return a string with the (printable) abbreviated storage index."""
def get_uri():
"""Return the (string) URI of the object that was checked."""
def is_healthy():
"""Return a boolean, True if the file/dir is fully healthy, False if
it is damaged in any way. Non-distributed LIT files always return
True."""
def is_recoverable():
"""Return a boolean, True if the file/dir can be recovered, False if
not. Unrecoverable files are obviously unhealthy. Non-distributed LIT
files always return True."""
# the following methods all return None for non-distributed LIT files
def get_happiness():
"""Return the happiness count of the file."""
def get_encoding_needed():
"""Return 'k', the number of shares required for recovery."""
def get_encoding_expected():
"""Return 'N', the number of total shares generated."""
def get_share_counter_good():
"""Return the number of distinct good shares that were found. For
mutable files, this counts shares for the 'best' version."""
def get_share_counter_wrong():
"""For mutable files, return the number of shares for versions other
than the 'best' one (which is defined as being the recoverable
version with the highest sequence number, then the highest roothash).
These are either leftover shares from an older version (perhaps on a
server that was offline when an update occurred), shares from an
unrecoverable newer version, or shares from an alternate current
version that results from an uncoordinated write collision. For a
healthy file, this will equal 0. For immutable files, this will
always equal 0."""
def get_corrupt_shares():
"""Return a list of 'share locators', one for each share that was
found to be corrupt (integrity failure). Each share locator is a list
of (IServer, storage_index, sharenum)."""
def get_incompatible_shares():
"""Return a list of 'share locators', one for each share that was
found to be of an unknown format. Each share locator is a list of
(IServer, storage_index, sharenum)."""
def get_servers_responding():
"""Return a list of IServer objects, one for each server that
responded to the share query (even if they said they didn't have
shares, and even if they said they did have shares but then didn't
send them when asked, or dropped the connection, or returned a
Failure, and even if they said they did have shares and sent
incorrect ones when asked)"""
def get_host_counter_good_shares():
"""Return the number of distinct storage servers with good shares. If
this number is less than get_share_counters()[good], then some shares
are doubled up, increasing the correlation of failures. This
indicates that one or more shares should be moved to an otherwise
unused server, if one is available.
"""
def get_version_counter_recoverable():
"""Return the number of recoverable versions of the file. For a
healthy file, this will equal 1."""
def get_version_counter_unrecoverable():
"""Return the number of unrecoverable versions of the file. For a
healthy file, this will be 0."""
def get_sharemap():
"""Return a dict mapping share identifier to list of IServer objects.
This indicates which servers are holding which shares. For immutable
files, the shareid is an integer (the share number, from 0 to N-1).
For mutable files, it is a string of the form 'seq%d-%s-sh%d',
containing the sequence number, the roothash, and the share number."""
def get_summary():
"""Return a string with a brief (one-line) summary of the results."""
def get_report():
"""Return a list of strings with more detailed results."""
class ICheckAndRepairResults(Interface):
"""I contain the detailed results of a check/verify/repair operation.
The IFilesystemNode.check()/verify()/repair() methods all return
instances that provide ICheckAndRepairResults.
"""
def get_storage_index():
"""Return a string with the (binary) storage index."""
def get_storage_index_string():
"""Return a string with the (printable) abbreviated storage index."""
def get_repair_attempted():
"""Return a boolean, True if a repair was attempted. We might not
attempt to repair the file because it was healthy, or healthy enough
(i.e. some shares were missing but not enough to exceed some
threshold), or because we don't know how to repair this object."""
def get_repair_successful():
"""Return a boolean, True if repair was attempted and the file/dir
was fully healthy afterwards. False if no repair was attempted or if
a repair attempt failed."""
def get_pre_repair_results():
"""Return an ICheckResults instance that describes the state of the
file/dir before any repair was attempted."""
def get_post_repair_results():
"""Return an ICheckResults instance that describes the state of the
file/dir after any repair was attempted. If no repair was attempted,
the pre-repair and post-repair results will be identical."""
class IDeepCheckResults(Interface):
"""I contain the results of a deep-check operation.
This is returned by a call to ICheckable.deep_check().
"""
def get_root_storage_index_string():
"""Return the storage index (abbreviated human-readable string) of
the first object checked."""
def get_counters():
"""Return a dictionary with the following keys::
count-objects-checked: count of how many objects were checked
count-objects-healthy: how many of those objects were completely
healthy
count-objects-unhealthy: how many were damaged in some way
count-objects-unrecoverable: how many were unrecoverable
count-corrupt-shares: how many shares were found to have
corruption, summed over all objects
examined
"""
def get_corrupt_shares():
"""Return a set of (IServer, storage_index, sharenum) for all shares
that were found to be corrupt. storage_index is binary."""
def get_all_results():
"""Return a dictionary mapping pathname (a tuple of strings, ready to
be slash-joined) to an ICheckResults instance, one for each object
that was checked."""
def get_results_for_storage_index(storage_index):
"""Retrive the ICheckResults instance for the given (binary)
storage index. Raises KeyError if there are no results for that
storage index."""
def get_stats():
"""Return a dictionary with the same keys as
IDirectoryNode.deep_stats()."""
class IDeepCheckAndRepairResults(Interface):
"""I contain the results of a deep-check-and-repair operation.
This is returned by a call to ICheckable.deep_check_and_repair().
"""
def get_root_storage_index_string():
"""Return the storage index (abbreviated human-readable string) of
the first object checked."""
def get_counters():
"""Return a dictionary with the following keys::
count-objects-checked: count of how many objects were checked
count-objects-healthy-pre-repair: how many of those objects were
completely healthy (before any
repair)
count-objects-unhealthy-pre-repair: how many were damaged in
some way
count-objects-unrecoverable-pre-repair: how many were unrecoverable
count-objects-healthy-post-repair: how many of those objects were
completely healthy (after any
repair)
count-objects-unhealthy-post-repair: how many were damaged in
some way
count-objects-unrecoverable-post-repair: how many were
unrecoverable
count-repairs-attempted: repairs were attempted on this many
objects. The count-repairs- keys will
always be provided, however unless
repair=true is present, they will all
be zero.
count-repairs-successful: how many repairs resulted in healthy
objects
count-repairs-unsuccessful: how many repairs resulted did not
results in completely healthy objects
count-corrupt-shares-pre-repair: how many shares were found to
have corruption, summed over all
objects examined (before any
repair)
count-corrupt-shares-post-repair: how many shares were found to
have corruption, summed over all
objects examined (after any
repair)
"""
def get_stats():
"""Return a dictionary with the same keys as
IDirectoryNode.deep_stats()."""
def get_corrupt_shares():
"""Return a set of (IServer, storage_index, sharenum) for all shares
that were found to be corrupt before any repair was attempted.
storage_index is binary.
"""
def get_remaining_corrupt_shares():
"""Return a set of (IServer, storage_index, sharenum) for all shares
that were found to be corrupt after any repair was completed.
storage_index is binary. These are shares that need manual inspection
and probably deletion.
"""
def get_all_results():
"""Return a dictionary mapping pathname (a tuple of strings, ready to
be slash-joined) to an ICheckAndRepairResults instance, one for each
object that was checked."""
def get_results_for_storage_index(storage_index):
"""Retrive the ICheckAndRepairResults instance for the given (binary)
storage index. Raises KeyError if there are no results for that
storage index."""
class IRepairable(Interface):
def repair(check_results):
"""Attempt to repair the given object. Returns a Deferred that fires
with a IRepairResults object.
I must be called with an object that implements ICheckResults, as
proof that you have actually discovered a problem with this file. I
will use the data in the checker results to guide the repair process,
such as which servers provided bad data and should therefore be
avoided. The ICheckResults object is inside the
ICheckAndRepairResults object, which is returned by the
ICheckable.check() method::
d = filenode.check(repair=False)
def _got_results(check_and_repair_results):
check_results = check_and_repair_results.get_pre_repair_results()
return filenode.repair(check_results)
d.addCallback(_got_results)
return d
"""
class IRepairResults(Interface):
"""I contain the results of a repair operation."""
def get_successful():
"""Returns a boolean: True if the repair made the file healthy, False
if not. Repair failure generally indicates a file that has been
damaged beyond repair."""
class IClient(Interface):
def upload(uploadable):
"""Upload some data into a CHK, get back the UploadResults for it.
@param uploadable: something that implements IUploadable
@return: a Deferred that fires with the UploadResults instance.
To get the URI for this file, use results.uri .
"""
def create_mutable_file(contents=""):
"""Create a new mutable file (with initial) contents, get back the
new node instance.
@param contents: (bytestring, callable, or None): this provides the
initial contents of the mutable file. If 'contents' is a bytestring,
it will be used as-is. If 'contents' is a callable, it will be
invoked with the new MutableFileNode instance and is expected to
return a bytestring with the initial contents of the file (the
callable can use node.get_writekey() to decide how to encrypt the
initial contents, e.g. for a brand new dirnode with initial
children). contents=None is equivalent to an empty string. Using
content_maker= is more efficient than creating a mutable file and
setting its contents in two separate operations.
@return: a Deferred that fires with an IMutableFileNode instance.
"""
def create_dirnode(initial_children={}):
"""Create a new unattached dirnode, possibly with initial children.
@param initial_children: dict with keys that are unicode child names,
and values that are (childnode, metadata) tuples.
@return: a Deferred that fires with the new IDirectoryNode instance.
"""
def create_node_from_uri(uri, rouri):
"""Create a new IFilesystemNode instance from the uri, synchronously.
@param uri: a string or IURI-providing instance, or None. This could
be for a LiteralFileNode, a CHK file node, a mutable file
node, or a directory node
@param rouri: a string or IURI-providing instance, or None. If the
main uri is None, I will use the rouri instead. If I
recognize the format of the main uri, I will ignore the
rouri (because it can be derived from the writecap).
@return: an instance that provides IFilesystemNode (or more usefully
one of its subclasses). File-specifying URIs will result in
IFileNode-providing instances, like ImmutableFileNode,
LiteralFileNode, or MutableFileNode. Directory-specifying
URIs will result in IDirectoryNode-providing instances, like
DirectoryNode.
"""
class INodeMaker(Interface):
"""The NodeMaker is used to create IFilesystemNode instances. It can
accept a filecap/dircap string and return the node right away. It can
also create new nodes (i.e. upload a file, or create a mutable file)
asynchronously. Once you have one of these nodes, you can use other
methods to determine whether it is a file or directory, and to download
or modify its contents.
The NodeMaker encapsulates all the authorities that these
IFilesystemNodes require (like references to the StorageFarmBroker). Each
Tahoe process will typically have a single NodeMaker, but unit tests may
create simplified/mocked forms for testing purposes.
"""
def create_from_cap(writecap, readcap=None, deep_immutable=False, name=u"<unknown name>"):
"""I create an IFilesystemNode from the given writecap/readcap. I can
only provide nodes for existing file/directory objects: use my other
methods to create new objects. I return synchronously."""
def create_mutable_file(contents=None, keysize=None):
"""I create a new mutable file, and return a Deferred that will fire
with the IMutableFileNode instance when it is ready. If contents= is
provided (a bytestring), it will be used as the initial contents of
the new file, otherwise the file will contain zero bytes. keysize= is
for use by unit tests, to create mutable files that are smaller than
usual."""
def create_new_mutable_directory(initial_children={}):
"""I create a new mutable directory, and return a Deferred that will
fire with the IDirectoryNode instance when it is ready. If
initial_children= is provided (a dict mapping unicode child name to
(childnode, metadata_dict) tuples), the directory will be populated
with those children, otherwise it will be empty."""
class IClientStatus(Interface):
def list_all_uploads():
"""Return a list of uploader objects, one for each upload that
currently has an object available (tracked with weakrefs). This is
intended for debugging purposes."""
def list_active_uploads():
"""Return a list of active IUploadStatus objects."""
def list_recent_uploads():
"""Return a list of IUploadStatus objects for the most recently
started uploads."""
def list_all_downloads():
"""Return a list of downloader objects, one for each download that
currently has an object available (tracked with weakrefs). This is
intended for debugging purposes."""
def list_active_downloads():
"""Return a list of active IDownloadStatus objects."""
def list_recent_downloads():
"""Return a list of IDownloadStatus objects for the most recently
started downloads."""
class IUploadStatus(Interface):
def get_started():
"""Return a timestamp (float with seconds since epoch) indicating
when the operation was started."""
def get_storage_index():
"""Return a string with the (binary) storage index in use on this
upload. Returns None if the storage index has not yet been
calculated."""
def get_size():
"""Return an integer with the number of bytes that will eventually
be uploaded for this file. Returns None if the size is not yet known.
"""
def using_helper():
"""Return True if this upload is using a Helper, False if not."""
def get_status():
"""Return a string describing the current state of the upload
process."""
def get_progress():
"""Returns a tuple of floats, (chk, ciphertext, encode_and_push),
each from 0.0 to 1.0 . 'chk' describes how much progress has been
made towards hashing the file to determine a CHK encryption key: if
non-convergent encryption is in use, this will be trivial, otherwise
the whole file must be hashed. 'ciphertext' describes how much of the
ciphertext has been pushed to the helper, and is '1.0' for non-helper
uploads. 'encode_and_push' describes how much of the encode-and-push
process has finished: for helper uploads this is dependent upon the
helper providing progress reports. It might be reasonable to add all
three numbers and report the sum to the user."""
def get_active():
"""Return True if the upload is currently active, False if not."""
def get_results():
"""Return an instance of UploadResults (which contains timing and
sharemap information). Might return None if the upload is not yet
finished."""
def get_counter():
"""Each upload status gets a unique number: this method returns that
number. This provides a handle to this particular upload, so a web
page can generate a suitable hyperlink."""
class IDownloadStatus(Interface):
def get_started():
"""Return a timestamp (float with seconds since epoch) indicating
when the operation was started."""
def get_storage_index():
"""Return a string with the (binary) storage index in use on this
download. This may be None if there is no storage index (i.e. LIT
files)."""
def get_size():
"""Return an integer with the number of bytes that will eventually be
retrieved for this file. Returns None if the size is not yet known.
"""
def using_helper():
"""Return True if this download is using a Helper, False if not."""
def get_status():
"""Return a string describing the current state of the download
process."""
def get_progress():
"""Returns a float (from 0.0 to 1.0) describing the amount of the
download that has completed. This value will remain at 0.0 until the
first byte of plaintext is pushed to the download target."""
def get_active():
"""Return True if the download is currently active, False if not."""
def get_counter():
"""Each download status gets a unique number: this method returns
that number. This provides a handle to this particular download, so a
web page can generate a suitable hyperlink."""
class IServermapUpdaterStatus(Interface):
pass
class IPublishStatus(Interface):
pass
class IRetrieveStatus(Interface):
pass
class NotCapableError(Exception):
"""You have tried to write to a read-only node."""
class BadWriteEnablerError(Exception):
pass
class RIControlClient(RemoteInterface):
def wait_for_client_connections(num_clients=int):
"""Do not return until we have connections to at least NUM_CLIENTS
storage servers.
"""
# debug stuff
def upload_random_data_from_file(size=int, convergence=str):
return str
def download_to_tempfile_and_delete(uri=str):
return None
def get_memory_usage():
"""Return a dict describes the amount of memory currently in use. The
keys are 'VmPeak', 'VmSize', and 'VmData'. The values are integers,
measuring memory consupmtion in bytes."""
return DictOf(str, int)
def speed_test(count=int, size=int, mutable=Any()):
"""Write 'count' tempfiles to disk, all of the given size. Measure
how long (in seconds) it takes to upload them all to the servers.
Then measure how long it takes to download all of them. If 'mutable'
is 'create', time creation of mutable files. If 'mutable' is
'upload', then time access to the same mutable file instead of
creating one.
Returns a tuple of (upload_time, download_time).
"""
return (float, float)
def measure_peer_response_time():
"""Send a short message to each connected peer, and measure the time
it takes for them to respond to it. This is a rough measure of the
application-level round trip time.
@return: a dictionary mapping peerid to a float (RTT time in seconds)
"""
return DictOf(str, float)
UploadResults = Any() #DictOf(str, str)
class RIEncryptedUploadable(RemoteInterface):
__remote_name__ = "RIEncryptedUploadable.tahoe.allmydata.com"
def get_size():
return Offset
def get_all_encoding_parameters():
return (int, int, int, long)
def read_encrypted(offset=Offset, length=ReadSize):
return ListOf(str)
def close():
return None
class RICHKUploadHelper(RemoteInterface):
__remote_name__ = "RIUploadHelper.tahoe.allmydata.com"
def get_version():
"""
Return a dictionary of version information.
"""
return DictOf(str, Any())
def upload(reader=RIEncryptedUploadable):
return UploadResults
class RIHelper(RemoteInterface):
__remote_name__ = "RIHelper.tahoe.allmydata.com"
def get_version():
"""
Return a dictionary of version information.
"""
return DictOf(str, Any())
def upload_chk(si=StorageIndex):
"""See if a file with a given storage index needs uploading. The
helper will ask the appropriate storage servers to see if the file
has already been uploaded. If so, the helper will return a set of
'upload results' that includes whatever hashes are needed to build
the read-cap, and perhaps a truncated sharemap.
If the file has not yet been uploaded (or if it was only partially
uploaded), the helper will return an empty upload-results dictionary
and also an RICHKUploadHelper object that will take care of the
upload process. The client should call upload() on this object and
pass it a reference to an RIEncryptedUploadable object that will
provide ciphertext. When the upload is finished, the upload() method
will finish and return the upload results.
"""
return (UploadResults, ChoiceOf(RICHKUploadHelper, None))
class RIStatsProvider(RemoteInterface):
__remote_name__ = "RIStatsProvider.tahoe.allmydata.com"
"""
Provides access to statistics and monitoring information.
"""
def get_stats():
"""
returns a dictionary containing 'counters' and 'stats', each a
dictionary with string counter/stat name keys, and numeric or None values.
counters are monotonically increasing measures of work done, and
stats are instantaneous measures (potentially time averaged
internally)
"""
return DictOf(str, DictOf(str, ChoiceOf(float, int, long, None)))
class RIStatsGatherer(RemoteInterface):
__remote_name__ = "RIStatsGatherer.tahoe.allmydata.com"
"""
Provides a monitoring service for centralised collection of stats
"""
def provide(provider=RIStatsProvider, nickname=str):
"""
@param provider: a stats collector instance that should be polled
periodically by the gatherer to collect stats.
@param nickname: a name useful to identify the provided client
"""
return None
class IStatsProducer(Interface):
def get_stats():
"""
returns a dictionary, with str keys representing the names of stats
to be monitored, and numeric values.
"""
class FileTooLargeError(Exception):
pass
class IValidatedThingProxy(Interface):
def start():
""" Acquire a thing and validate it. Return a deferred that is
eventually fired with self if the thing is valid or errbacked if it
can't be acquired or validated."""
class InsufficientVersionError(Exception):
def __init__(self, needed, got):
self.needed = needed
self.got = got
def __repr__(self):
return "InsufficientVersionError(need '%s', got %s)" % (self.needed,
self.got)
class EmptyPathnameComponentError(Exception):
"""The webapi disallows empty pathname components."""
Reference in New Issue View Git Blame Copy Permalink