Madiator2011/lbry-sdk
1
0
Fork 0
forked from LBRYCommunity/lbry-sdk
Code Pull requests Activity

Merge pull request #3631 from lbryio/bootstrap_node

Browse source 
Add all peers when running as a bootstrap node
This commit is contained in:
Victor Shyba 2022-08-29 11:07:29 -03:00 committed by GitHub
commit a334a93757
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
8 changed files with 165 additions and 164 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
lbry/conf.py
View file

@ -624,6 +624,10 @@ class Config(CLIConfig):
"will increase. This setting is used by seed nodes, you probably don't want to change it during normal " "will increase. This setting is used by seed nodes, you probably don't want to change it during normal "
"use.", 2 "use.", 2
) )
is_bootstrap_node = Toggle(
"When running as a bootstrap node, disable all logic related to balancing the routing table, so we can "
"add as many peers as possible and better help first-runs.", False
)
# protocol timeouts # protocol timeouts
download_timeout = Float("Cumulative timeout for a stream to begin downloading before giving up", 30.0) download_timeout = Float("Cumulative timeout for a stream to begin downloading before giving up", 30.0)

21
lbry/dht/node.py
View file

@ -30,12 +30,12 @@ class Node:
) )
def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', node_id: bytes, udp_port: int, def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', node_id: bytes, udp_port: int,
internal_udp_port: int, peer_port: int, external_ip: str, rpc_timeout: float = constants.RPC_TIMEOUT, internal_udp_port: int, peer_port: int, external_ip: str, rpc_timeout: float = constants.RPC_TIMEOUT,
split_buckets_under_index: int = constants.SPLIT_BUCKETS_UNDER_INDEX, split_buckets_under_index: int = constants.SPLIT_BUCKETS_UNDER_INDEX, is_bootstrap_node: bool = False,
storage: typing.Optional['SQLiteStorage'] = None): storage: typing.Optional['SQLiteStorage'] = None):
self.loop = loop self.loop = loop
self.internal_udp_port = internal_udp_port self.internal_udp_port = internal_udp_port
self.protocol = KademliaProtocol(loop, peer_manager, node_id, external_ip, udp_port, peer_port, rpc_timeout, self.protocol = KademliaProtocol(loop, peer_manager, node_id, external_ip, udp_port, peer_port, rpc_timeout,
split_buckets_under_index) split_buckets_under_index, is_bootstrap_node)
self.listening_port: asyncio.DatagramTransport = None self.listening_port: asyncio.DatagramTransport = None
self.joined = asyncio.Event(loop=self.loop) self.joined = asyncio.Event(loop=self.loop)
self._join_task: asyncio.Task = None self._join_task: asyncio.Task = None
@ -70,13 +70,6 @@ class Node:
# get ids falling in the midpoint of each bucket that hasn't been recently updated # get ids falling in the midpoint of each bucket that hasn't been recently updated
node_ids = self.protocol.routing_table.get_refresh_list(0, True) node_ids = self.protocol.routing_table.get_refresh_list(0, True)
# if we have 3 or fewer populated buckets get two random ids in the range of each to try and
# populate/split the buckets further
buckets_with_contacts = self.protocol.routing_table.buckets_with_contacts()
if buckets_with_contacts <= 3:
for i in range(buckets_with_contacts):
node_ids.append(self.protocol.routing_table.random_id_in_bucket_range(i))
node_ids.append(self.protocol.routing_table.random_id_in_bucket_range(i))
if self.protocol.routing_table.get_peers(): if self.protocol.routing_table.get_peers():
# if we have node ids to look up, perform the iterative search until we have k results # if we have node ids to look up, perform the iterative search until we have k results
@ -203,15 +196,13 @@ class Node:
def get_iterative_node_finder(self, key: bytes, shortlist: typing.Optional[typing.List['KademliaPeer']] = None, def get_iterative_node_finder(self, key: bytes, shortlist: typing.Optional[typing.List['KademliaPeer']] = None,
max_results: int = constants.K) -> IterativeNodeFinder: max_results: int = constants.K) -> IterativeNodeFinder:
shortlist = shortlist or self.protocol.routing_table.find_close_peers(key)
return IterativeNodeFinder(self.loop, self.protocol.peer_manager, self.protocol.routing_table, self.protocol, return IterativeNodeFinder(self.loop, self.protocol, key, max_results, shortlist)
key, max_results, None, shortlist)
def get_iterative_value_finder(self, key: bytes, shortlist: typing.Optional[typing.List['KademliaPeer']] = None, def get_iterative_value_finder(self, key: bytes, shortlist: typing.Optional[typing.List['KademliaPeer']] = None,
max_results: int = -1) -> IterativeValueFinder: max_results: int = -1) -> IterativeValueFinder:
shortlist = shortlist or self.protocol.routing_table.find_close_peers(key)
return IterativeValueFinder(self.loop, self.protocol.peer_manager, self.protocol.routing_table, self.protocol, return IterativeValueFinder(self.loop, self.protocol, key, max_results, shortlist)
key, max_results, None, shortlist)
async def peer_search(self, node_id: bytes, count=constants.K, max_results=constants.K * 2, async def peer_search(self, node_id: bytes, count=constants.K, max_results=constants.K * 2,
shortlist: typing.Optional[typing.List['KademliaPeer']] = None shortlist: typing.Optional[typing.List['KademliaPeer']] = None

45
lbry/dht/protocol/iterative_find.py
View file

@ -12,7 +12,6 @@ from lbry.dht.peer import make_kademlia_peer
from lbry.dht.serialization.datagram import PAGE_KEY from lbry.dht.serialization.datagram import PAGE_KEY
if TYPE_CHECKING: if TYPE_CHECKING:
from lbry.dht.protocol.routing_table import TreeRoutingTable
from lbry.dht.protocol.protocol import KademliaProtocol from lbry.dht.protocol.protocol import KademliaProtocol
from lbry.dht.peer import PeerManager, KademliaPeer from lbry.dht.peer import PeerManager, KademliaPeer
@ -57,37 +56,19 @@ class FindValueResponse(FindResponse):
return [(node_id, address.decode(), port) for node_id, address, port in self.close_triples] return [(node_id, address.decode(), port) for node_id, address, port in self.close_triples]
def get_shortlist(routing_table: 'TreeRoutingTable', key: bytes,
shortlist: typing.Optional[typing.List['KademliaPeer']]) -> typing.List['KademliaPeer']:
"""
If not provided, initialize the shortlist of peers to probe to the (up to) k closest peers in the routing table
:param routing_table: a TreeRoutingTable
:param key: a 48 byte hash
:param shortlist: optional manually provided shortlist, this is done during bootstrapping when there are no
peers in the routing table. During bootstrap the shortlist is set to be the seed nodes.
"""
if len(key) != constants.HASH_LENGTH:
raise ValueError("invalid key length: %i" % len(key))
return shortlist or routing_table.find_close_peers(key)
class IterativeFinder(AsyncIterator): class IterativeFinder(AsyncIterator):
def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', def __init__(self, loop: asyncio.AbstractEventLoop,
routing_table: 'TreeRoutingTable', protocol: 'KademliaProtocol', key: bytes, protocol: 'KademliaProtocol', key: bytes,
max_results: typing.Optional[int] = constants.K, max_results: typing.Optional[int] = constants.K,
exclude: typing.Optional[typing.List[typing.Tuple[str, int]]] = None,
shortlist: typing.Optional[typing.List['KademliaPeer']] = None): shortlist: typing.Optional[typing.List['KademliaPeer']] = None):
if len(key) != constants.HASH_LENGTH: if len(key) != constants.HASH_LENGTH:
raise ValueError("invalid key length: %i" % len(key)) raise ValueError("invalid key length: %i" % len(key))
self.loop = loop self.loop = loop
self.peer_manager = peer_manager self.peer_manager = protocol.peer_manager
self.routing_table = routing_table
self.protocol = protocol self.protocol = protocol
self.key = key self.key = key
self.max_results = max(constants.K, max_results) self.max_results = max(constants.K, max_results)
self.exclude = exclude or []
self.active: typing.Dict['KademliaPeer', int] = OrderedDict() # peer: distance, sorted self.active: typing.Dict['KademliaPeer', int] = OrderedDict() # peer: distance, sorted
self.contacted: typing.Set['KademliaPeer'] = set() self.contacted: typing.Set['KademliaPeer'] = set()
@ -99,7 +80,7 @@ class IterativeFinder(AsyncIterator):
self.iteration_count = 0 self.iteration_count = 0
self.running = False self.running = False
self.tasks: typing.List[asyncio.Task] = [] self.tasks: typing.List[asyncio.Task] = []
for peer in get_shortlist(routing_table, key, shortlist): for peer in shortlist:
if peer.node_id: if peer.node_id:
self._add_active(peer, force=True) self._add_active(peer, force=True)
else: else:
@ -198,8 +179,6 @@ class IterativeFinder(AsyncIterator):
if index > (constants.K + len(self.running_probes)): if index > (constants.K + len(self.running_probes)):
break break
origin_address = (peer.address, peer.udp_port) origin_address = (peer.address, peer.udp_port)
if origin_address in self.exclude:
continue
if peer.node_id == self.protocol.node_id: if peer.node_id == self.protocol.node_id:
continue continue
if origin_address == (self.protocol.external_ip, self.protocol.udp_port): if origin_address == (self.protocol.external_ip, self.protocol.udp_port):
@ -277,13 +256,11 @@ class IterativeFinder(AsyncIterator):
type(self).__name__, id(self), self.key.hex()[:8]) type(self).__name__, id(self), self.key.hex()[:8])
class IterativeNodeFinder(IterativeFinder): class IterativeNodeFinder(IterativeFinder):
def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', def __init__(self, loop: asyncio.AbstractEventLoop,
routing_table: 'TreeRoutingTable', protocol: 'KademliaProtocol', key: bytes, protocol: 'KademliaProtocol', key: bytes,
max_results: typing.Optional[int] = constants.K, max_results: typing.Optional[int] = constants.K,
exclude: typing.Optional[typing.List[typing.Tuple[str, int]]] = None,
shortlist: typing.Optional[typing.List['KademliaPeer']] = None): shortlist: typing.Optional[typing.List['KademliaPeer']] = None):
super().__init__(loop, peer_manager, routing_table, protocol, key, max_results, exclude, super().__init__(loop, protocol, key, max_results, shortlist)
shortlist)
self.yielded_peers: typing.Set['KademliaPeer'] = set() self.yielded_peers: typing.Set['KademliaPeer'] = set()
async def send_probe(self, peer: 'KademliaPeer') -> FindNodeResponse: async def send_probe(self, peer: 'KademliaPeer') -> FindNodeResponse:
@ -319,13 +296,11 @@ class IterativeNodeFinder(IterativeFinder):
class IterativeValueFinder(IterativeFinder): class IterativeValueFinder(IterativeFinder):
def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', def __init__(self, loop: asyncio.AbstractEventLoop,
routing_table: 'TreeRoutingTable', protocol: 'KademliaProtocol', key: bytes, protocol: 'KademliaProtocol', key: bytes,
max_results: typing.Optional[int] = constants.K, max_results: typing.Optional[int] = constants.K,
exclude: typing.Optional[typing.List[typing.Tuple[str, int]]] = None,
shortlist: typing.Optional[typing.List['KademliaPeer']] = None): shortlist: typing.Optional[typing.List['KademliaPeer']] = None):
super().__init__(loop, peer_manager, routing_table, protocol, key, max_results, exclude, super().__init__(loop, protocol, key, max_results, shortlist)
shortlist)
self.blob_peers: typing.Set['KademliaPeer'] = set() self.blob_peers: typing.Set['KademliaPeer'] = set()
# this tracks the index of the most recent page we requested from each peer # this tracks the index of the most recent page we requested from each peer
self.peer_pages: typing.DefaultDict['KademliaPeer', int] = defaultdict(int) self.peer_pages: typing.DefaultDict['KademliaPeer', int] = defaultdict(int)

87
lbry/dht/protocol/protocol.py
View file

@ -218,6 +218,10 @@ class PingQueue:
def running(self): def running(self):
return self._running return self._running
@property
def busy(self):
return self._running and (any(self._running_pings) or any(self._pending_contacts))
def enqueue_maybe_ping(self, *peers: 'KademliaPeer', delay: typing.Optional[float] = None): def enqueue_maybe_ping(self, *peers: 'KademliaPeer', delay: typing.Optional[float] = None):
delay = delay if delay is not None else self._default_delay delay = delay if delay is not None else self._default_delay
now = self._loop.time() now = self._loop.time()
@ -229,7 +233,7 @@ class PingQueue:
async def ping_task(): async def ping_task():
try: try:
if self._protocol.peer_manager.peer_is_good(peer): if self._protocol.peer_manager.peer_is_good(peer):
if peer not in self._protocol.routing_table.get_peers(): if not self._protocol.routing_table.get_peer(peer.node_id):
self._protocol.add_peer(peer) self._protocol.add_peer(peer)
return return
await self._protocol.get_rpc_peer(peer).ping() await self._protocol.get_rpc_peer(peer).ping()
@ -294,7 +298,7 @@ class KademliaProtocol(DatagramProtocol):
def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', node_id: bytes, external_ip: str, def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', node_id: bytes, external_ip: str,
udp_port: int, peer_port: int, rpc_timeout: float = constants.RPC_TIMEOUT, udp_port: int, peer_port: int, rpc_timeout: float = constants.RPC_TIMEOUT,
split_buckets_under_index: int = constants.SPLIT_BUCKETS_UNDER_INDEX): split_buckets_under_index: int = constants.SPLIT_BUCKETS_UNDER_INDEX, is_boostrap_node: bool = False):
self.peer_manager = peer_manager self.peer_manager = peer_manager
self.loop = loop self.loop = loop
self.node_id = node_id self.node_id = node_id
@ -309,7 +313,8 @@ class KademliaProtocol(DatagramProtocol):
self.transport: DatagramTransport = None self.transport: DatagramTransport = None
self.old_token_secret = constants.generate_id() self.old_token_secret = constants.generate_id()
self.token_secret = constants.generate_id() self.token_secret = constants.generate_id()
self.routing_table = TreeRoutingTable(self.loop, self.peer_manager, self.node_id, split_buckets_under_index) self.routing_table = TreeRoutingTable(
self.loop, self.peer_manager, self.node_id, split_buckets_under_index, is_bootstrap_node=is_boostrap_node)
self.data_store = DictDataStore(self.loop, self.peer_manager) self.data_store = DictDataStore(self.loop, self.peer_manager)
self.ping_queue = PingQueue(self.loop, self) self.ping_queue = PingQueue(self.loop, self)
self.node_rpc = KademliaRPC(self, self.loop, self.peer_port) self.node_rpc = KademliaRPC(self, self.loop, self.peer_port)
@ -356,72 +361,10 @@ class KademliaProtocol(DatagramProtocol):
return args, {} return args, {}
async def _add_peer(self, peer: 'KademliaPeer'): async def _add_peer(self, peer: 'KademliaPeer'):
if not peer.node_id: async def probe(some_peer: 'KademliaPeer'):
log.warning("Tried adding a peer with no node id!") rpc_peer = self.get_rpc_peer(some_peer)
return False await rpc_peer.ping()
for my_peer in self.routing_table.get_peers(): return await self.routing_table.add_peer(peer, probe)
if (my_peer.address, my_peer.udp_port) == (peer.address, peer.udp_port) and my_peer.node_id != peer.node_id:
self.routing_table.remove_peer(my_peer)
self.routing_table.join_buckets()
bucket_index = self.routing_table.kbucket_index(peer.node_id)
if self.routing_table.buckets[bucket_index].add_peer(peer):
return True
# The bucket is full; see if it can be split (by checking if its range includes the host node's node_id)
if self.routing_table.should_split(bucket_index, peer.node_id):
self.routing_table.split_bucket(bucket_index)
# Retry the insertion attempt
result = await self._add_peer(peer)
self.routing_table.join_buckets()
return result
else:
# We can't split the k-bucket
#
# The 13 page kademlia paper specifies that the least recently contacted node in the bucket
# shall be pinged. If it fails to reply it is replaced with the new contact. If the ping is successful
# the new contact is ignored and not added to the bucket (sections 2.2 and 2.4).
#
# A reasonable extension to this is BEP 0005, which extends the above:
#
# Not all nodes that we learn about are equal. Some are "good" and some are not.
# Many nodes using the DHT are able to send queries and receive responses,
# but are not able to respond to queries from other nodes. It is important that
# each node's routing table must contain only known good nodes. A good node is
# a node has responded to one of our queries within the last 15 minutes. A node
# is also good if it has ever responded to one of our queries and has sent us a
# query within the last 15 minutes. After 15 minutes of inactivity, a node becomes
# questionable. Nodes become bad when they fail to respond to multiple queries
# in a row. Nodes that we know are good are given priority over nodes with unknown status.
#
# When there are bad or questionable nodes in the bucket, the least recent is selected for
# potential replacement (BEP 0005). When all nodes in the bucket are fresh, the head (least recent)
# contact is selected as described in section 2.2 of the kademlia paper. In both cases the new contact
# is ignored if the pinged node replies.
not_good_contacts = self.routing_table.buckets[bucket_index].get_bad_or_unknown_peers()
not_recently_replied = []
for my_peer in not_good_contacts:
last_replied = self.peer_manager.get_last_replied(my_peer.address, my_peer.udp_port)
if not last_replied or last_replied + 60 < self.loop.time():
not_recently_replied.append(my_peer)
if not_recently_replied:
to_replace = not_recently_replied[0]
else:
to_replace = self.routing_table.buckets[bucket_index].peers[0]
last_replied = self.peer_manager.get_last_replied(to_replace.address, to_replace.udp_port)
if last_replied and last_replied + 60 > self.loop.time():
return False
log.debug("pinging %s:%s", to_replace.address, to_replace.udp_port)
try:
to_replace_rpc = self.get_rpc_peer(to_replace)
await to_replace_rpc.ping()
return False
except asyncio.TimeoutError:
log.debug("Replacing dead contact in bucket %i: %s:%i with %s:%i ", bucket_index,
to_replace.address, to_replace.udp_port, peer.address, peer.udp_port)
if to_replace in self.routing_table.buckets[bucket_index]:
self.routing_table.buckets[bucket_index].remove_peer(to_replace)
return await self._add_peer(peer)
def add_peer(self, peer: 'KademliaPeer'): def add_peer(self, peer: 'KademliaPeer'):
if peer.node_id == self.node_id: if peer.node_id == self.node_id:
@ -439,7 +382,6 @@ class KademliaProtocol(DatagramProtocol):
async with self._split_lock: async with self._split_lock:
peer = self._to_remove.pop() peer = self._to_remove.pop()
self.routing_table.remove_peer(peer) self.routing_table.remove_peer(peer)
self.routing_table.join_buckets()
while self._to_add: while self._to_add:
async with self._split_lock: async with self._split_lock:
await self._add_peer(self._to_add.pop()) await self._add_peer(self._to_add.pop())
@ -482,9 +424,8 @@ class KademliaProtocol(DatagramProtocol):
# This is an RPC method request # This is an RPC method request
self.received_request_metric.labels(method=request_datagram.method).inc() self.received_request_metric.labels(method=request_datagram.method).inc()
self.peer_manager.report_last_requested(address[0], address[1]) self.peer_manager.report_last_requested(address[0], address[1])
try: peer = self.routing_table.get_peer(request_datagram.node_id)
peer = self.routing_table.get_peer(request_datagram.node_id) if not peer:
except IndexError:
try: try:
peer = make_kademlia_peer(request_datagram.node_id, address[0], address[1]) peer = make_kademlia_peer(request_datagram.node_id, address[0], address[1])
except ValueError as err: except ValueError as err:

140
lbry/dht/protocol/routing_table.py
View file

@ -28,7 +28,8 @@ class KBucket:
namespace="dht_node", labelnames=("amount",) namespace="dht_node", labelnames=("amount",)
) )
def __init__(self, peer_manager: 'PeerManager', range_min: int, range_max: int, node_id: bytes): def __init__(self, peer_manager: 'PeerManager', range_min: int, range_max: int,
node_id: bytes, capacity: int = constants.K):
""" """
@param range_min: The lower boundary for the range in the n-bit ID @param range_min: The lower boundary for the range in the n-bit ID
space covered by this k-bucket space covered by this k-bucket
@ -36,12 +37,12 @@ class KBucket:
covered by this k-bucket covered by this k-bucket
""" """
self._peer_manager = peer_manager self._peer_manager = peer_manager
self.last_accessed = 0
self.range_min = range_min self.range_min = range_min
self.range_max = range_max self.range_max = range_max
self.peers: typing.List['KademliaPeer'] = [] self.peers: typing.List['KademliaPeer'] = []
self._node_id = node_id self._node_id = node_id
self._distance_to_self = Distance(node_id) self._distance_to_self = Distance(node_id)
self.capacity = capacity
def add_peer(self, peer: 'KademliaPeer') -> bool: def add_peer(self, peer: 'KademliaPeer') -> bool:
""" Add contact to _contact list in the right order. This will move the """ Add contact to _contact list in the right order. This will move the
@ -68,7 +69,7 @@ class KBucket:
self.peers.remove(local_peer) self.peers.remove(local_peer)
self.peers.append(peer) self.peers.append(peer)
return True return True
if len(self.peers) < constants.K: if len(self.peers) < self.capacity:
self.peers.append(peer) self.peers.append(peer)
self.peer_in_routing_table_metric.labels("global").inc() self.peer_in_routing_table_metric.labels("global").inc()
bits_colliding = utils.get_colliding_prefix_bits(peer.node_id, self._node_id) bits_colliding = utils.get_colliding_prefix_bits(peer.node_id, self._node_id)
@ -76,13 +77,11 @@ class KBucket:
return True return True
else: else:
return False return False
# raise BucketFull("No space in bucket to insert contact")
def get_peer(self, node_id: bytes) -> 'KademliaPeer': def get_peer(self, node_id: bytes) -> 'KademliaPeer':
for peer in self.peers: for peer in self.peers:
if peer.node_id == node_id: if peer.node_id == node_id:
return peer return peer
raise IndexError(node_id)
def get_peers(self, count=-1, exclude_contact=None, sort_distance_to=None) -> typing.List['KademliaPeer']: def get_peers(self, count=-1, exclude_contact=None, sort_distance_to=None) -> typing.List['KademliaPeer']:
""" Returns a list containing up to the first count number of contacts """ Returns a list containing up to the first count number of contacts
@ -179,6 +178,13 @@ class TreeRoutingTable:
version of the Kademlia paper, in section 2.4. It does, however, use the version of the Kademlia paper, in section 2.4. It does, however, use the
ping RPC-based k-bucket eviction algorithm described in section 2.2 of ping RPC-based k-bucket eviction algorithm described in section 2.2 of
that paper. that paper.
BOOTSTRAP MODE: if set to True, we always add all peers. This is so a
bootstrap node does not get a bias towards its own node id and replies are
the best it can provide (joining peer knows its neighbors immediately).
Over time, this will need to be optimized so we use the disk as holding
everything in memory won't be feasible anymore.
See: https://github.com/bittorrent/bootstrap-dht
""" """
bucket_in_routing_table_metric = Gauge( bucket_in_routing_table_metric = Gauge(
"buckets_in_routing_table", "Number of buckets on routing table", namespace="dht_node", "buckets_in_routing_table", "Number of buckets on routing table", namespace="dht_node",
@ -186,21 +192,22 @@ class TreeRoutingTable:
) )
def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', parent_node_id: bytes, def __init__(self, loop: asyncio.AbstractEventLoop, peer_manager: 'PeerManager', parent_node_id: bytes,
split_buckets_under_index: int = constants.SPLIT_BUCKETS_UNDER_INDEX): split_buckets_under_index: int = constants.SPLIT_BUCKETS_UNDER_INDEX, is_bootstrap_node: bool = False):
self._loop = loop self._loop = loop
self._peer_manager = peer_manager self._peer_manager = peer_manager
self._parent_node_id = parent_node_id self._parent_node_id = parent_node_id
self._split_buckets_under_index = split_buckets_under_index self._split_buckets_under_index = split_buckets_under_index
self.buckets: typing.List[KBucket] = [ self.buckets: typing.List[KBucket] = [
KBucket( KBucket(
self._peer_manager, range_min=0, range_max=2 ** constants.HASH_BITS, node_id=self._parent_node_id self._peer_manager, range_min=0, range_max=2 ** constants.HASH_BITS, node_id=self._parent_node_id,
capacity=1 << 32 if is_bootstrap_node else constants.K
) )
] ]
def get_peers(self) -> typing.List['KademliaPeer']: def get_peers(self) -> typing.List['KademliaPeer']:
return list(itertools.chain.from_iterable(map(lambda bucket: bucket.peers, self.buckets))) return list(itertools.chain.from_iterable(map(lambda bucket: bucket.peers, self.buckets)))
def should_split(self, bucket_index: int, to_add: bytes) -> bool: def _should_split(self, bucket_index: int, to_add: bytes) -> bool:
# https://stackoverflow.com/questions/32129978/highly-unbalanced-kademlia-routing-table/32187456#32187456 # https://stackoverflow.com/questions/32129978/highly-unbalanced-kademlia-routing-table/32187456#32187456
if bucket_index < self._split_buckets_under_index: if bucket_index < self._split_buckets_under_index:
return True return True
@ -225,39 +232,32 @@ class TreeRoutingTable:
return [] return []
def get_peer(self, contact_id: bytes) -> 'KademliaPeer': def get_peer(self, contact_id: bytes) -> 'KademliaPeer':
""" return self.buckets[self._kbucket_index(contact_id)].get_peer(contact_id)
@raise IndexError: No contact with the specified contact ID is known
by this node
"""
return self.buckets[self.kbucket_index(contact_id)].get_peer(contact_id)
def get_refresh_list(self, start_index: int = 0, force: bool = False) -> typing.List[bytes]: def get_refresh_list(self, start_index: int = 0, force: bool = False) -> typing.List[bytes]:
bucket_index = start_index
refresh_ids = [] refresh_ids = []
now = int(self._loop.time()) for offset, _ in enumerate(self.buckets[start_index:]):
for bucket in self.buckets[start_index:]: refresh_ids.append(self._midpoint_id_in_bucket_range(start_index + offset))
if force or now - bucket.last_accessed >= constants.REFRESH_INTERVAL: # if we have 3 or fewer populated buckets get two random ids in the range of each to try and
to_search = self.midpoint_id_in_bucket_range(bucket_index) # populate/split the buckets further
refresh_ids.append(to_search) buckets_with_contacts = self.buckets_with_contacts()
bucket_index += 1 if buckets_with_contacts <= 3:
for i in range(buckets_with_contacts):
refresh_ids.append(self._random_id_in_bucket_range(i))
refresh_ids.append(self._random_id_in_bucket_range(i))
return refresh_ids return refresh_ids
def remove_peer(self, peer: 'KademliaPeer') -> None: def remove_peer(self, peer: 'KademliaPeer') -> None:
if not peer.node_id: if not peer.node_id:
return return
bucket_index = self.kbucket_index(peer.node_id) bucket_index = self._kbucket_index(peer.node_id)
try: try:
self.buckets[bucket_index].remove_peer(peer) self.buckets[bucket_index].remove_peer(peer)
self._join_buckets()
except ValueError: except ValueError:
return return
def touch_kbucket(self, key: bytes) -> None: def _kbucket_index(self, key: bytes) -> int:
self.touch_kbucket_by_index(self.kbucket_index(key))
def touch_kbucket_by_index(self, bucket_index: int):
self.buckets[bucket_index].last_accessed = int(self._loop.time())
def kbucket_index(self, key: bytes) -> int:
i = 0 i = 0
for bucket in self.buckets: for bucket in self.buckets:
if bucket.key_in_range(key): if bucket.key_in_range(key):
@ -266,19 +266,19 @@ class TreeRoutingTable:
i += 1 i += 1
return i return i
def random_id_in_bucket_range(self, bucket_index: int) -> bytes: def _random_id_in_bucket_range(self, bucket_index: int) -> bytes:
random_id = int(random.randrange(self.buckets[bucket_index].range_min, self.buckets[bucket_index].range_max)) random_id = int(random.randrange(self.buckets[bucket_index].range_min, self.buckets[bucket_index].range_max))
return Distance( return Distance(
self._parent_node_id self._parent_node_id
)(random_id.to_bytes(constants.HASH_LENGTH, 'big')).to_bytes(constants.HASH_LENGTH, 'big') )(random_id.to_bytes(constants.HASH_LENGTH, 'big')).to_bytes(constants.HASH_LENGTH, 'big')
def midpoint_id_in_bucket_range(self, bucket_index: int) -> bytes: def _midpoint_id_in_bucket_range(self, bucket_index: int) -> bytes:
half = int((self.buckets[bucket_index].range_max - self.buckets[bucket_index].range_min) // 2) half = int((self.buckets[bucket_index].range_max - self.buckets[bucket_index].range_min) // 2)
return Distance(self._parent_node_id)( return Distance(self._parent_node_id)(
int(self.buckets[bucket_index].range_min + half).to_bytes(constants.HASH_LENGTH, 'big') int(self.buckets[bucket_index].range_min + half).to_bytes(constants.HASH_LENGTH, 'big')
).to_bytes(constants.HASH_LENGTH, 'big') ).to_bytes(constants.HASH_LENGTH, 'big')
def split_bucket(self, old_bucket_index: int) -> None: def _split_bucket(self, old_bucket_index: int) -> None:
""" Splits the specified k-bucket into two new buckets which together """ Splits the specified k-bucket into two new buckets which together
cover the same range in the key/ID space cover the same range in the key/ID space
@ -303,7 +303,7 @@ class TreeRoutingTable:
old_bucket.remove_peer(contact) old_bucket.remove_peer(contact)
self.bucket_in_routing_table_metric.labels("global").set(len(self.buckets)) self.bucket_in_routing_table_metric.labels("global").set(len(self.buckets))
def join_buckets(self): def _join_buckets(self):
if len(self.buckets) == 1: if len(self.buckets) == 1:
return return
to_pop = [i for i, bucket in enumerate(self.buckets) if len(bucket) == 0] to_pop = [i for i, bucket in enumerate(self.buckets) if len(bucket) == 0]
@ -326,14 +326,7 @@ class TreeRoutingTable:
self.buckets[bucket_index_to_pop + 1].range_min = bucket.range_min self.buckets[bucket_index_to_pop + 1].range_min = bucket.range_min
self.buckets.remove(bucket) self.buckets.remove(bucket)
self.bucket_in_routing_table_metric.labels("global").set(len(self.buckets)) self.bucket_in_routing_table_metric.labels("global").set(len(self.buckets))
return self.join_buckets() return self._join_buckets()
def contact_in_routing_table(self, address_tuple: typing.Tuple[str, int]) -> bool:
for bucket in self.buckets:
for contact in bucket.get_peers(sort_distance_to=False):
if address_tuple[0] == contact.address and address_tuple[1] == contact.udp_port:
return True
return False
def buckets_with_contacts(self) -> int: def buckets_with_contacts(self) -> int:
count = 0 count = 0
@ -341,3 +334,70 @@ class TreeRoutingTable:
if len(bucket) > 0: if len(bucket) > 0:
count += 1 count += 1
return count return count
async def add_peer(self, peer: 'KademliaPeer', probe: typing.Callable[['KademliaPeer'], typing.Awaitable]):
if not peer.node_id:
log.warning("Tried adding a peer with no node id!")
return False
for my_peer in self.get_peers():
if (my_peer.address, my_peer.udp_port) == (peer.address, peer.udp_port) and my_peer.node_id != peer.node_id:
self.remove_peer(my_peer)
self._join_buckets()
bucket_index = self._kbucket_index(peer.node_id)
if self.buckets[bucket_index].add_peer(peer):
return True
# The bucket is full; see if it can be split (by checking if its range includes the host node's node_id)
if self._should_split(bucket_index, peer.node_id):
self._split_bucket(bucket_index)
# Retry the insertion attempt
result = await self.add_peer(peer, probe)
self._join_buckets()
return result
else:
# We can't split the k-bucket
#
# The 13 page kademlia paper specifies that the least recently contacted node in the bucket
# shall be pinged. If it fails to reply it is replaced with the new contact. If the ping is successful
# the new contact is ignored and not added to the bucket (sections 2.2 and 2.4).
#
# A reasonable extension to this is BEP 0005, which extends the above:
#
# Not all nodes that we learn about are equal. Some are "good" and some are not.
# Many nodes using the DHT are able to send queries and receive responses,
# but are not able to respond to queries from other nodes. It is important that
# each node's routing table must contain only known good nodes. A good node is
# a node has responded to one of our queries within the last 15 minutes. A node
# is also good if it has ever responded to one of our queries and has sent us a
# query within the last 15 minutes. After 15 minutes of inactivity, a node becomes
# questionable. Nodes become bad when they fail to respond to multiple queries
# in a row. Nodes that we know are good are given priority over nodes with unknown status.
#
# When there are bad or questionable nodes in the bucket, the least recent is selected for
# potential replacement (BEP 0005). When all nodes in the bucket are fresh, the head (least recent)
# contact is selected as described in section 2.2 of the kademlia paper. In both cases the new contact
# is ignored if the pinged node replies.
not_good_contacts = self.buckets[bucket_index].get_bad_or_unknown_peers()
not_recently_replied = []
for my_peer in not_good_contacts:
last_replied = self._peer_manager.get_last_replied(my_peer.address, my_peer.udp_port)
if not last_replied or last_replied + 60 < self._loop.time():
not_recently_replied.append(my_peer)
if not_recently_replied:
to_replace = not_recently_replied[0]
else:
to_replace = self.buckets[bucket_index].peers[0]
last_replied = self._peer_manager.get_last_replied(to_replace.address, to_replace.udp_port)
if last_replied and last_replied + 60 > self._loop.time():
return False
log.debug("pinging %s:%s", to_replace.address, to_replace.udp_port)
try:
await probe(to_replace)
return False
except asyncio.TimeoutError:
log.debug("Replacing dead contact in bucket %i: %s:%i with %s:%i ", bucket_index,
to_replace.address, to_replace.udp_port, peer.address, peer.udp_port)
if to_replace in self.buckets[bucket_index]:
self.buckets[bucket_index].remove_peer(to_replace)
return await self.add_peer(peer, probe)

1
lbry/extras/daemon/components.py
View file

@ -297,6 +297,7 @@ class DHTComponent(Component):
peer_port=self.external_peer_port, peer_port=self.external_peer_port,
rpc_timeout=self.conf.node_rpc_timeout, rpc_timeout=self.conf.node_rpc_timeout,
split_buckets_under_index=self.conf.split_buckets_under_index, split_buckets_under_index=self.conf.split_buckets_under_index,
is_bootstrap_node=self.conf.is_bootstrap_node,
storage=storage storage=storage
) )
self.dht_node.start(self.conf.network_interface, self.conf.known_dht_nodes) self.dht_node.start(self.conf.network_interface, self.conf.known_dht_nodes)

2
scripts/dht_node.py
View file

@ -83,7 +83,7 @@ async def main(host: str, port: int, db_file_path: str, bootstrap_node: Optional
await storage.open() await storage.open()
node = Node( node = Node(
loop, PeerManager(loop), node_id, port, port, 3333, None, loop, PeerManager(loop), node_id, port, port, 3333, None,
storage=storage storage=storage, is_bootstrap_node=True
) )
if prometheus_port > 0: if prometheus_port > 0:
metrics = SimpleMetrics(prometheus_port, node if export else None) metrics = SimpleMetrics(prometheus_port, node if export else None)

29
tests/unit/dht/test_node.py
View file

@ -11,6 +11,35 @@ from lbry.dht.peer import PeerManager, make_kademlia_peer
from lbry.extras.daemon.storage import SQLiteStorage from lbry.extras.daemon.storage import SQLiteStorage
class TestBootstrapNode(AsyncioTestCase):
TIMEOUT = 10.0 # do not increase. Hitting a timeout is a real failure
async def test_bootstrap_node_adds_all_peers(self):
loop = asyncio.get_event_loop()
loop.set_debug(False)
with dht_mocks.mock_network_loop(loop):
advance = dht_mocks.get_time_accelerator(loop)
self.bootstrap_node = Node(self.loop, PeerManager(loop), constants.generate_id(),
4444, 4444, 3333, '1.2.3.4', is_bootstrap_node=True)
self.bootstrap_node.start('1.2.3.4', [])
self.bootstrap_node.protocol.ping_queue._default_delay = 0
self.addCleanup(self.bootstrap_node.stop)
# start the nodes
nodes = {}
futs = []
for i in range(100):
nodes[i] = Node(loop, PeerManager(loop), constants.generate_id(i), 4444, 4444, 3333, f'1.3.3.{i}')
nodes[i].start(f'1.3.3.{i}', [('1.2.3.4', 4444)])
self.addCleanup(nodes[i].stop)
futs.append(nodes[i].joined.wait())
await asyncio.gather(*futs)
while self.bootstrap_node.protocol.ping_queue.busy:
await advance(1)
self.assertEqual(100, len(self.bootstrap_node.protocol.routing_table.get_peers()))
class TestNodePingQueueDiscover(AsyncioTestCase): class TestNodePingQueueDiscover(AsyncioTestCase):
async def test_ping_queue_discover(self): async def test_ping_queue_discover(self):
loop = asyncio.get_event_loop() loop = asyncio.get_event_loop()