Browse Source

lnrouter fixes:

- use gossip_queries_req instead of initial_routing_sync
 - add connected nodes to recent peers only after successful init
 - derive timestamp used with gossip_timestamp_filter from channel_db
 - fix query_short_channel_ids:
     1. channel IDs must be sorted with zlib
     2. limit request to 100
     3. do not abuse this to request node_announcements; it is fine not to have all nodes.
 - fix get_recent_peers:
     1. do not set last_connected_date to 'now' if we never connected a node
     2. sql query was misconstructed and was returning only one peer
 - populate FALLBACK_NODE_LIST_MAINNET with nodes that have the requested flags
regtest_lnd
ThomasV 6 years ago
committed by SomberNight
parent
commit
b624de41e9
No known key found for this signature in database GPG Key ID: B33B5F232C6271E9
  1. 70
      electrum/lnpeer.py
  2. 56
      electrum/lnrouter.py
  3. 41
      electrum/lnworker.py
  4. 6
      electrum/tests/test_lnpeer.py

70
electrum/lnpeer.py

@ -14,6 +14,7 @@ from functools import partial
from typing import List, Tuple, Dict, TYPE_CHECKING, Optional, Callable
import traceback
import sys
from datetime import datetime
import aiorpcx
@ -53,8 +54,7 @@ def channel_id_from_funding_tx(funding_txid: str, funding_index: int) -> Tuple[b
class Peer(PrintError):
def __init__(self, lnworker: 'LNWorker', pubkey:bytes, transport: LNTransportBase,
request_initial_sync=False):
def __init__(self, lnworker: 'LNWorker', pubkey:bytes, transport: LNTransportBase):
self.initialized = asyncio.Event()
self.node_anns = []
self.chan_anns = []
@ -77,8 +77,7 @@ class Peer(PrintError):
self.closing_signed = defaultdict(asyncio.Queue)
self.payment_preimages = defaultdict(asyncio.Queue)
self.localfeatures = LnLocalFeatures(0)
if request_initial_sync:
self.localfeatures |= LnLocalFeatures.INITIAL_ROUTING_SYNC
self.localfeatures |= LnLocalFeatures.GOSSIP_QUERIES_REQ
self.localfeatures |= LnLocalFeatures.OPTION_DATA_LOSS_PROTECT_REQ
self.attempted_route = {}
self.orphan_channel_updates = OrderedDict()
@ -96,7 +95,6 @@ class Peer(PrintError):
async def initialize(self):
if isinstance(self.transport, LNTransport):
await self.transport.handshake()
self.channel_db.add_recent_peer(self.transport.peer_addr)
self.send_message("init", gflen=0, lflen=1, localfeatures=self.localfeatures)
@property
@ -172,8 +170,8 @@ class Peer(PrintError):
raise LightningPeerConnectionClosed("remote does not have even flag {}"
.format(str(LnLocalFeatures(1 << flag))))
self.localfeatures ^= 1 << flag # disable flag
first_timestamp = self.lnworker.get_first_timestamp()
self.send_message('gossip_timestamp_filter', chain_hash=constants.net.rev_genesis_bytes(), first_timestamp=first_timestamp, timestamp_range=b"\xff"*4)
if isinstance(self.transport, LNTransport):
self.channel_db.add_recent_peer(self.transport.peer_addr)
self.initialized.set()
def on_node_announcement(self, payload):
@ -215,6 +213,17 @@ class Peer(PrintError):
@log_exceptions
async def _gossip_loop(self):
await self.initialized.wait()
timestamp = self.channel_db.get_last_timestamp()
if timestamp == 0:
self.print_error('requesting whole channel graph')
else:
self.print_error('requesting channel graph since', datetime.fromtimestamp(timestamp).ctime())
timestamp_range = int(time.time()) - timestamp
self.send_message(
'gossip_timestamp_filter',
chain_hash=constants.net.rev_genesis_bytes(),
first_timestamp=timestamp,
timestamp_range=timestamp_range)
while True:
await asyncio.sleep(5)
if self.node_anns:
@ -226,13 +235,13 @@ class Peer(PrintError):
if self.chan_upds:
self.channel_db.on_channel_update(self.chan_upds)
self.chan_upds = []
need_to_get = self.channel_db.missing_short_chan_ids() #type: Set[int]
need_to_get = sorted(self.channel_db.missing_short_chan_ids())
if need_to_get and not self.receiving_channels:
self.print_error('QUERYING SHORT CHANNEL IDS; missing', len(need_to_get), 'channels')
zlibencoded = zlib.compress(bfh(''.join(need_to_get)))
self.print_error('missing', len(need_to_get), 'channels')
zlibencoded = zlib.compress(bfh(''.join(need_to_get[0:100])))
self.send_message(
'query_short_channel_ids',
chain_hash=bytes.fromhex(bitcoin.rev_hex(constants.net.GENESIS)),
chain_hash=constants.net.rev_genesis_bytes(),
len=1+len(zlibencoded),
encoded_short_ids=b'\x01' + zlibencoded)
self.receiving_channels = True
@ -705,20 +714,33 @@ class Peer(PrintError):
# that the remote sends, even if the channel was not announced
# (from BOLT-07: "MAY create a channel_update to communicate the channel
# parameters to the final node, even though the channel has not yet been announced")
self.channel_db.on_channel_announcement({"short_channel_id": chan.short_channel_id, "node_id_1": node_ids[0], "node_id_2": node_ids[1],
'chain_hash': constants.net.rev_genesis_bytes(), 'len': b'\x00\x00', 'features': b'',
'bitcoin_key_1': bitcoin_keys[0], 'bitcoin_key_2': bitcoin_keys[1]},
trusted=True)
self.channel_db.on_channel_announcement(
{
"short_channel_id": chan.short_channel_id,
"node_id_1": node_ids[0],
"node_id_2": node_ids[1],
'chain_hash': constants.net.rev_genesis_bytes(),
'len': b'\x00\x00',
'features': b'',
'bitcoin_key_1': bitcoin_keys[0],
'bitcoin_key_2': bitcoin_keys[1]
},
trusted=True)
# only inject outgoing direction:
if node_ids[0] == privkey_to_pubkey(self.privkey):
channel_flags = b'\x00'
else:
channel_flags = b'\x01'
now = int(time.time()).to_bytes(4, byteorder="big")
self.channel_db.on_channel_update({"short_channel_id": chan.short_channel_id, 'channel_flags': channel_flags, 'cltv_expiry_delta': b'\x90',
'htlc_minimum_msat': b'\x03\xe8', 'fee_base_msat': b'\x03\xe8', 'fee_proportional_millionths': b'\x01',
'chain_hash': constants.net.rev_genesis_bytes(), 'timestamp': now},
trusted=True)
channel_flags = b'\x00' if node_ids[0] == privkey_to_pubkey(self.privkey) else b'\x01'
now = int(time.time())
self.channel_db.on_channel_update(
{
"short_channel_id": chan.short_channel_id,
'channel_flags': channel_flags,
'cltv_expiry_delta': b'\x90',
'htlc_minimum_msat': b'\x03\xe8',
'fee_base_msat': b'\x03\xe8',
'fee_proportional_millionths': b'\x01',
'chain_hash': constants.net.rev_genesis_bytes(),
'timestamp': now.to_bytes(4, byteorder="big")
},
trusted=True)
# peer may have sent us a channel update for the incoming direction previously
# note: if we were offline when the 3rd conf happened, lnd will never send us this channel_update
# see https://github.com/lightningnetwork/lnd/issues/1347

56
electrum/lnrouter.py

@ -23,6 +23,7 @@
# CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.
from datetime import datetime
import time
import random
import queue
@ -157,9 +158,8 @@ class NodeInfo(Base):
addresses = NodeInfo.parse_addresses_field(payload['addresses'])
alias = payload['alias'].rstrip(b'\x00').hex()
timestamp = int.from_bytes(payload['timestamp'], "big")
now = int(time.time())
return NodeInfo(node_id=node_id, features=features, timestamp=timestamp, alias=alias), [
Address(host=host, port=port, node_id=node_id, last_connected_date=now) for host, port in addresses]
Address(host=host, port=port, node_id=node_id, last_connected_date=None) for host, port in addresses]
@staticmethod
def parse_addresses_field(addresses_field):
@ -206,8 +206,7 @@ class Address(Base):
node_id = Column(String(66), ForeignKey('node_info.node_id'), primary_key=True)
host = Column(String(256), primary_key=True)
port = Column(Integer, primary_key=True)
last_connected_date = Column(Integer(), nullable=False)
last_connected_date = Column(Integer(), nullable=True)
@ -273,11 +272,8 @@ class ChannelDB(SqlDB):
@sql
def get_recent_peers(self):
return [LNPeerAddr(x.host, x.port, bytes.fromhex(x.node_id)) for x in self.DBSession \
.query(Address) \
.select_from(NodeInfo) \
.order_by(Address.last_connected_date.desc()) \
.limit(self.NUM_MAX_RECENT_PEERS)]
r = self.DBSession.query(Address).filter(Address.last_connected_date.isnot(None)).order_by(Address.last_connected_date.desc()).limit(self.NUM_MAX_RECENT_PEERS).all()
return [LNPeerAddr(x.host, x.port, bytes.fromhex(x.node_id)) for x in r]
@sql
def get_channel_info(self, channel_id: bytes):
@ -298,15 +294,6 @@ class ChannelDB(SqlDB):
chan_ids_from_policy = set(x[0] for x in self.DBSession.query(Policy.short_channel_id).filter(expr).all())
if chan_ids_from_policy:
return chan_ids_from_policy
# fetch channels for node_ids missing in node_info. that will also give us node_announcement
expr = not_(ChannelInfo.node1_id.in_(self.DBSession.query(NodeInfo.node_id)))
chan_ids_from_id1 = set(x[0] for x in self.DBSession.query(ChannelInfo.short_channel_id).filter(expr).all())
if chan_ids_from_id1:
return chan_ids_from_id1
expr = not_(ChannelInfo.node2_id.in_(self.DBSession.query(NodeInfo.node_id)))
chan_ids_from_id2 = set(x[0] for x in self.DBSession.query(ChannelInfo.short_channel_id).filter(expr).all())
if chan_ids_from_id2:
return chan_ids_from_id2
return set()
@sql
@ -318,7 +305,7 @@ class ChannelDB(SqlDB):
self.DBSession.commit()
@sql
#@profiler
@profiler
def on_channel_announcement(self, msg_payloads, trusted=False):
if type(msg_payloads) is dict:
msg_payloads = [msg_payloads]
@ -342,10 +329,16 @@ class ChannelDB(SqlDB):
for channel_info in new_channels.values():
self.DBSession.add(channel_info)
self.DBSession.commit()
self.print_error('on_channel_announcement: %d/%d'%(len(new_channels), len(msg_payloads)))
#self.print_error('on_channel_announcement: %d/%d'%(len(new_channels), len(msg_payloads)))
self._update_counts()
self.network.trigger_callback('ln_status')
@sql
def get_last_timestamp(self):
from sqlalchemy.sql import func
r = self.DBSession.query(func.max(Policy.timestamp).label('max_timestamp')).one()
return r.max_timestamp or 0
@sql
@profiler
def on_channel_update(self, msg_payloads, trusted=False):
@ -368,7 +361,8 @@ class ChannelDB(SqlDB):
if not trusted and not verify_sig_for_channel_update(msg_payload, bytes.fromhex(node_id)):
continue
short_channel_id = channel_info.short_channel_id
new_policy = Policy.from_msg(msg_payload, node_id, channel_info.short_channel_id)
new_policy = Policy.from_msg(msg_payload, node_id, short_channel_id)
#self.print_error('on_channel_update', datetime.fromtimestamp(new_policy.timestamp).ctime())
old_policy = self.DBSession.query(Policy).filter_by(short_channel_id=short_channel_id, start_node=node_id).one_or_none()
if old_policy:
if old_policy.timestamp >= new_policy.timestamp:
@ -378,6 +372,7 @@ class ChannelDB(SqlDB):
if p and p.timestamp >= new_policy.timestamp:
continue
new_policies[(short_channel_id, node_id)] = new_policy
self.print_error('on_channel_update: %d/%d'%(len(new_policies), len(msg_payloads)))
# commit pending removals
self.DBSession.commit()
# add and commit new policies
@ -386,7 +381,7 @@ class ChannelDB(SqlDB):
self.DBSession.commit()
@sql
#@profiler
@profiler
def on_node_announcement(self, msg_payloads):
if type(msg_payloads) is dict:
msg_payloads = [msg_payloads]
@ -403,7 +398,13 @@ class ChannelDB(SqlDB):
node_info, node_addresses = NodeInfo.from_msg(msg_payload)
except UnknownEvenFeatureBits:
continue
#self.print_error('received node announcement from', datetime.fromtimestamp(node_info.timestamp).ctime())
node_id = node_info.node_id
# Ignore node if it has no associated channel (DoS protection)
expr = or_(ChannelInfo.node1_id==node_id, ChannelInfo.node2_id==node_id)
if self.DBSession.query(ChannelInfo.short_channel_id).filter(expr).count() == 0:
#self.print_error('ignoring orphan node_announcement')
continue
node = self.DBSession.query(NodeInfo).filter_by(node_id=node_id).one_or_none()
if node and node.timestamp >= node_info.timestamp:
continue
@ -413,20 +414,13 @@ class ChannelDB(SqlDB):
new_nodes[node_id] = node_info
for addr in node_addresses:
new_addresses[(addr.node_id,addr.host,addr.port)] = addr
self.print_error("on_node_announcements: %d/%d"%(len(new_nodes), len(msg_payloads)))
self.print_error("on_node_announcement: %d/%d"%(len(new_nodes), len(msg_payloads)))
for node_info in new_nodes.values():
self.DBSession.add(node_info)
for new_addr in new_addresses.values():
old_addr = self.DBSession.query(Address).filter_by(node_id=new_addr.node_id, host=new_addr.host, port=new_addr.port).one_or_none()
if old_addr:
old_addr.last_connected_date = new_addr.last_connected_date
else:
if not old_addr:
self.DBSession.add(new_addr)
# TODO if this message is for a new node, and if we have no associated
# channels for this node, we should ignore the message and return here,
# to mitigate DOS. but race condition: the channels we have for this
# node, might be under verification in self.ca_verifier, what then?
self.DBSession.commit()
self._update_counts()
self.network.trigger_callback('ln_status')

41
electrum/lnworker.py

@ -59,10 +59,16 @@ FALLBACK_NODE_LIST_TESTNET = (
LNPeerAddr('148.251.87.112', 9735, bfh('021a8bd8d8f1f2e208992a2eb755cdc74d44e66b6a0c924d3a3cce949123b9ce40')), # janus test server
LNPeerAddr('122.199.61.90', 9735, bfh('038863cf8ab91046230f561cd5b386cbff8309fa02e3f0c3ed161a3aeb64a643b9')), # popular node https://1ml.com/testnet/node/038863cf8ab91046230f561cd5b386cbff8309fa02e3f0c3ed161a3aeb64a643b9
)
FALLBACK_NODE_LIST_MAINNET = (
LNPeerAddr('104.198.32.198', 9735, bfh('02f6725f9c1c40333b67faea92fd211c183050f28df32cac3f9d69685fe9665432')), # Blockstream
LNPeerAddr('13.80.67.162', 9735, bfh('02c0ac82c33971de096d87ce5ed9b022c2de678f08002dc37fdb1b6886d12234b5')), # Stampery
)
FALLBACK_NODE_LIST_MAINNET = [
LNPeerAddr(host='52.168.166.221', port=9735, pubkey=b'\x02\x148+\xdc\xe7u\r\xfc\xb8\x12m\xf8\xe2\xb1-\xe3\x856\x90-\xc3j\xbc\xeb\xda\xee\xfd\xec\xa1\xdf\x82\x84'),
LNPeerAddr(host='35.230.100.60', port=9735, pubkey=b'\x02?^5\x82qk\xed\x96\xf6\xf2l\xfc\xd8\x03~\x07GM{GC\xaf\xdc\x8b\x07\xe6\x92\xdfcFM~'),
LNPeerAddr(host='40.69.71.114', port=9735, pubkey=b'\x02\x83\x03\x18,\x98\x85\xda\x93\xb3\xb2\\\x96!\xd2,\xf3Du\xe6<\x129B\xe4\x02\xabS\x0c\x05V\xe6u'),
LNPeerAddr(host='62.210.110.5', port=9735, pubkey=b'\x02v\xe0\x9a&u\x92\xe7E\x1a\x93\x9c\x93,\xf6\x85\xf0uM\xe3\x82\xa3\xca\x85\xd2\xfb:\x86ML6Z\xd5'),
LNPeerAddr(host='34.236.113.58', port=9735, pubkey=b'\x02\xfaP\xc7.\xe1\xe2\xeb_\x1bm\x9c02\x08\x0cL\x86Cs\xc4 \x1d\xfa)f\xaa4\xee\xe1\x05\x1f\x97'),
LNPeerAddr(host='52.168.166.221', port=9735, pubkey=b'\x02\x148+\xdc\xe7u\r\xfc\xb8\x12m\xf8\xe2\xb1-\xe3\x856\x90-\xc3j\xbc\xeb\xda\xee\xfd\xec\xa1\xdf\x82\x84'),
LNPeerAddr(host='34.236.113.58', port=9735, pubkey=b'\x02\xfaP\xc7.\xe1\xe2\xeb_\x1bm\x9c02\x08\x0cL\x86Cs\xc4 \x1d\xfa)f\xaa4\xee\xe1\x05\x1f\x97'),
]
encoder = ChannelJsonEncoder()
@ -103,29 +109,6 @@ class LNWorker(PrintError):
asyncio.run_coroutine_threadsafe(self.network.main_taskgroup.spawn(self.main_loop()), self.network.asyncio_loop)
self.first_timestamp_requested = None
def get_first_timestamp(self):
first_request = False
if self.first_timestamp_requested is None:
self.first_timestamp_requested = time.time()
first_request = True
first_timestamp = self.storage.get('lightning_gossip_until', 0)
if first_timestamp == 0:
self.print_error('requesting whole channel graph')
else:
self.print_error('requesting channel graph since', datetime.fromtimestamp(first_timestamp).ctime())
if first_request:
asyncio.run_coroutine_threadsafe(self.save_gossip_timestamp(), self.network.asyncio_loop)
return first_timestamp
@log_exceptions
async def save_gossip_timestamp(self):
while True:
await asyncio.sleep(GRAPH_DOWNLOAD_SECONDS)
yesterday = int(time.time()) - 24*60*60 # now minus a day
self.storage.put('lightning_gossip_until', yesterday)
self.storage.write()
self.print_error('saved lightning gossip timestamp')
def payment_completed(self, chan: Channel, direction: Direction,
htlc: UpdateAddHtlc):
chan_id = chan.channel_id
@ -258,7 +241,7 @@ class LNWorker(PrintError):
transport = LNTransport(self.node_keypair.privkey, peer_addr)
self._last_tried_peer[peer_addr] = time.time()
self.print_error("adding peer", peer_addr)
peer = Peer(self, node_id, transport, request_initial_sync=self.config.get("request_initial_sync", True))
peer = Peer(self, node_id, transport)
await self.network.main_taskgroup.spawn(peer.main_loop())
self.peers[node_id] = peer
self.network.trigger_callback('ln_status')
@ -839,7 +822,7 @@ class LNWorker(PrintError):
except:
self.print_error('handshake failure from incoming connection')
return
peer = Peer(self, node_id, transport, request_initial_sync=self.config.get("request_initial_sync", True))
peer = Peer(self, node_id, transport)
self.peers[node_id] = peer
await self.network.main_taskgroup.spawn(peer.main_loop())
self.network.trigger_callback('ln_status')

6
electrum/tests/test_lnpeer.py

@ -178,7 +178,7 @@ class TestPeer(SequentialTestCase):
def test_require_data_loss_protect(self):
mock_lnworker = MockLNWorker(keypair(), keypair(), self.alice_channel, tx_queue=None)
mock_transport = NoFeaturesTransport('')
p1 = Peer(mock_lnworker, b"\x00" * 33, mock_transport, request_initial_sync=False)
p1 = Peer(mock_lnworker, b"\x00" * 33, mock_transport)
mock_lnworker.peer = p1
with self.assertRaises(LightningPeerConnectionClosed):
run(asyncio.wait_for(p1._message_loop(), 1))
@ -189,8 +189,8 @@ class TestPeer(SequentialTestCase):
q1, q2 = asyncio.Queue(), asyncio.Queue()
w1 = MockLNWorker(k1, k2, self.alice_channel, tx_queue=q1)
w2 = MockLNWorker(k2, k1, self.bob_channel, tx_queue=q2)
p1 = Peer(w1, k1.pubkey, t1, request_initial_sync=False)
p2 = Peer(w2, k2.pubkey, t2, request_initial_sync=False)
p1 = Peer(w1, k1.pubkey, t1)
p2 = Peer(w2, k2.pubkey, t2)
w1.peer = p1
w2.peer = p2
# mark_open won't work if state is already OPEN.

Loading…
Cancel
Save