This commit loosens the fwdpkg reference acking to be more tolerant
of prior deletions. Specifically, we won't fail if certain channels
are not found or fwdpkgs do not exist. This will make us more
tolerant to future changes where we:
- remove fwdpkgs on channel close
- defensively cleanup stray responses
In this commit, we add a new test to expose a lurking bug within the
graph database code. As is, when we go to delete a node from the
database, we don't also remove the entries within the update index. As a
result, if a user attempted to call NodeUpdatesInHorizon (or typically
as part of the p2p handshake), we would error out, as we would try to
read a node that no longer existed in the graph, as it was pruned.
In this commit we fix a minor logging artifact. After the switch to
EdgePoint, the FilteredChainView implementations will try to log the
struct directly, as prior they would have an outpoint object. We restore
this behavior by adding a String() method to EdgePoint which will simply
proxy through to the outpoint so we can log that directly.
In this commit, we add a new method to the ChannelEdgeInfo that will
allow the path finding logic to get the node opposite the pivot node
without first creating a new db transaction. The new method is able to
use an existing db transaction, or create a new one if needed.
The commit ensures that for every channel, there will always
be two entries in the edges bucket. If the policy from one or
both ends of the channel is unknown, it is marked as such.
This allows efficient lookup of incoming edges. This is
required for backwards payment path finding.
In this commit, we update the ChannelView method to be compatible with
the new set of interfaces that require the script to be passed in in
addition to the outpoint. In order to do this, we introduce a new
EdgePoint struct which packages together a channel point along with the
funding pkScript. Along the way, we've copied over a utility method from
the lnwallet package to avoid having to deal with an import cycle.
In this commit, we fix a slight race condition that can occur when we go
to add a shell node for a node announcement, but then right afterwards,
a new block arrives that causes us to prune an unconnected node. To
ensure this doesn't happen, we now add shell nodes within the same db
transaction as AddChannelEdge. This ensures that the state is fully
consistent and shell nodes will be added atomically along with the new
channel edge.
As a result of this change, we no longer need to add shell nodes within
the ChannelRouter, as the database will take care of this operation as
it should.
In this commit, we fix an existing bug in the pruneGraphNodes method.
Before this commit, if a node was involved in a channel, but only one of
the edges was advertised, then either it, or the other node would be
erroneously pruned from the graph. They shouldn't be pruned as there's
still an edge connecting the, although only 1/2 of the edge is actually
advertised.
In order to fix this, we'll now do two passes: the first pass will
populate a ref count map of all known nodes in the graph, the second
pass will increment the ref count each time a node is found in the
graph. With this two pass method, we ensure that nodes are only deleted
if there are absolutely no edges pointing to them within the graph.
In this commit, we extend the TestPruneGraphNodes test to also test the
case of when a node is involved in a channel, but only a single edge for
that channel has been advertised. In order to test this, we add an
additional node to the graph, and also a new channel. However, this
channel will only have a single edge advertised. As result, when we
prune the set of edges, the only node remaining should be the node that
didn't have any edges at all.
In this commit, we extend the server's functionality to prune link nodes
on startup. Since we currently only decide whether to prune a link node
from the database based on a channel close, it's possible that we have
link nodes lingering from before this functionality was added on.
In this commit, we fix an existing bug related to duplicate invoice
settle.s Before this commit, the second (and later) times an invoice was
settled we would return a nil pointer. This would result in the new
invoiceRegistry panicing as it would go to attempt to notify with a nil
invoice.
We fix this by returning the invoice on disk (unmodified) for each
settle after the initial one.
Fixes#1568.
In this commit, we add a new test to ensure that duplicate invoice
settles work as expected. At the present time, this test will fail as
the second to last assertion fails as we'll return a nil invoice the
second time around.
In this commit, we migrate the database away from a partially migrated
state. In a prior commit, we migrated the database in order to update
the Invoice struct with three new fields: add index, settle index, paid
amt. However, it was overlooked that the OutgoingPayment struct also
embedded an Invoice within it. As a result, nodes that upgraded to the
first migration found themselves unable to start up, or call
listpayments, as the internal invoice within the OutgoignPayment hadn't
yet been updated. This would result in an OOM typically as we went to
allocate a slice with a integer that should have been small, but may
have ended up actually being a set of random bytes, so a very large
number.
In this commit, we finish the DB migration by also migrating the
internal invoice within each OutgoingPayment.
Fixes#1538.
Fixes#1546.
In this commit, we fix an existing bu gin the invoice time series
migration code. Before this commit, the migration would fail as we would
try to migrate an empty invoice. We now detect this case and skip all
empty invoices.
We also add a bit more logging on both the info and trace logging level.
In this commit, we add two new methods: InvoicesAddedSince and
InvoicesSettledSince. These methods will be used by higher level
sub-systems that implement notifications to deliver any notifications
backlog based on the last add index, and last settle index that the
client knows of.
It's important to note that care has been taken to ensure that this new
API can be used in a backwards compatible manner. If a client specifies
and index of 0 for either of the methods, then no backlog will be sent.
This is due to the fact that current users of the API don't expect any
backlog notifications to be sent. Additionally, the index actually
starts at 1, instead of 0.
In this commit, we add two new indexes to the invoice database: the add
index, and the settle index. These to indexes essentially form a time
series index on top of the existing primary index bucket. Each time an
invoice is added, we'll advance the addIndex seqno, and then create a
mapping from seqNo -> invoiceNum. Each time an invoice is settled, we'll
do the same, but within the settle index.
This change is required in order to allow callers to effectively seek
into the current invoice database in order to obtain notifications for
any invoices they may have missed out on while they were disconnected.
This will allow us to implement robust streaming invoice notifications
within lnd to ensure that clients never miss an event.
In this commit, in order to allow the caller to specify the amount that
was ultimately accepted for an invoice, the SettleInvoice method has
gained a new parameter: amtPaid. SettleInvoice will now populate the
final amount paid in the database upon db commit.
In this commit, we move to explicitly storing a bit more information
within the invoice. Currently this information is already stored in the
payment request, but by storing it at this level, callers that may not
be in the state to fully decode a payment request can obtain this data.
We avoid a database migration by appending this data to the end of an
invoice. When decoding, we'll try to read out this extra information,
and simply return what we have if it isn't found.
This commit synchronizes the in-memory cache with the
on-disk state to ensure the waiting proof store is
externally consistent. Currently, there are scenarios
where the in-memory state is updated, and not reverted
if the write fails. The general fix is to wait to apply
modifications until the write succeeds, and use a
read/write lock to synchronize access with db operations.
In this commit, we fix an existing bug in the new graph query sync
feature. Before this commit, when a block is pruned, we would never
actually delete the update index entries. This is due to the fact that
we would attempt to delete the entries from the update index _after_ we
had already removed the edges from the update index.
We fix this by simply swapping the order: first we delete from the
update index, then we delete the edges themselves. A test ensuring that
the entires are cleared (which failed before this commit), has been
added.
In this commit, we go through the codebase looking for TCP address
assumptions and modifying them to include the recently introduced onion
addresses. This enables us to fully support onion addresses within the
daemon.
In this commit, we modify the waiting proof slightly to acept dupliacte
waiting proofs, rather than reject them. Otherwise, it's possible that
the remote node first sends us their half of the waiting proof (before
we do), we write that to disk, then upon restart, we'll try to add it
again, but be rejected by the system.
Fixes#1315.
In this commit, we ensure that all indexes for a particular channel have
any relevant keys deleted once a channel is removed from the database.
Before this commit, if we pruned a channel due to closing, then its
entry in the channel update index would ever be removed.
In this commit, we add a new database migration required to update old
database to the version of the database that tracks the update index for
the nodes and edge policies. The migration is straight forward, we
simply need to populate the new indexes for the all the nodes, and then
all the edges.
In this commit, we add a series of methods, and a new database index
that we'll use to implement the new discovery.ChannelGraphTimeSeries
interface interface. The primary change is that we now maintain two new
indexes tracking the last update time for each node, and the last update
time for each edge. These two indexes allow us to implement the
NodeUpdatesInHorizon and ChanUpdatesInHorizon methods. The remaining
methods added simply utilize the existing database indexes to allow us to
respond to any peer gossip range queries.
A set of new unit tests has been added to exercise the added logic.
The pending state definitin in ChannelCloseSummary was slightly changed
in such a way that channels that has had their commitment broadcasted
now is no longer considered "pending close". They now instead stay in
the open chan bucket with the ChanStatus "CommitmentBroadcasted" until
their commitment is confirmed. This commit updates the IsPending godoc
to reflect this.
In this commit, we modify the existing updateChanBucket function to no
longer auto-create buckets if they don't exist. We do this in order to
fix a class of bug that could arise wherein after a channel has actually
be closed (and the parent buckets removed) a method that mutates the
channel state is called, which then re-creates the relevant set of
buckets. As a result, subsequent calls to any RPCs which need to read
all the channels will fail as most of the fields won't actually be
populated.
After this commit, the fullSync method is the only one that's able to
create the full bucket hierarchy.
In this commit, we extend the CloseChannelSummary by also storing: the
current unrevoked revocation for the remote party, the next pending
unused revocation, and also the local channel config. We move to store
these as the provide an extra level of defense against bugs as we'll
always store information required to derive keys for any current and
prior states.
This commit adds a new method FetchWaitingCloseChannels to the database,
used for fetching OpenChannels that have a ChanStatus != Default. These
are channels that are borked, or have had a commitment broadcasted, and
is now waiting for it to confirm.
The fetchChannels method is rewritten to return channels exclusively
based on wheter they are pending or waitingClose.
This commit changes the bool `IsBorked` in OpenChannel to a `ChanStatus`
struct, of type ChannelStatus. This is used to indicated that a channel
that is technically still open, is either borked, or has had a
commitment broadcasted, but is not confirmed on-chain yet.
The ChannelStatus type has the value 1 for the status Borked, meaning it
is backwards compatible with the old database format.
Modifies TestFetchPendingChannels to verify that calls to
MarkAsOpen also modify the in-memory state. Previously we
only tested the persistent state loaded immediately after.
Modifies the MarkAsOpen operation to also update the
ShortChanID and IsPending fields in-memory. Before,
only the on-disk representation was updated, which
may have lead to stale data channel states being
passed in-memory.
In this commit, we remove references to raw keys from the main
ChannelConfig struct and instead replace it with usage of
keychain.KeyDescriptor. We do this, as the ChannelConfig as it stands
is a near complete static description of a channel. In the future, it
will be possible to export these static descriptions as backups. We
prefer the KeyDescriptor of a plain PublicKey, as the KeyLocator
portion of the struct allows a stateless signer to re-derive the keys
as needed when signing.
In this commit, we add a new storage namespace to channeldb: the
ForwardingLog. This log will be used by higher level sub-systems to log
each successfully completed HTLC. Each payment circuit will be
summarized as a “ForwardingEvent”. A series of events can then be
queried via a time slice query. In a time slice query, the caller
specifies a time range, a number of events to skip, and the max number
of events to return. Each query will return the index of the final
item. As we have a max number of events we’ll return in a response,
callers may need to use this last offset index to seek further by
skipping that number of entries. Combining these fields, callers are
able to query the time series, skipping an arbitrary amount of events,
and capping the max number of returned events.
In this commit, we made a series of modification to the way we handle
reading edges and vertexes from disk, in order to reduce the amount of
garbage generated:
1. Properly use PubKeyBytes are required rather than PubKey()
2. Return direct structs rather than pointers, and leave it to the
runtime to perform escape analysis.
3. In-line the former readSig() method when reading sigs from disk.
This commit changes the definition of the
constraints in the ChannelConstraints struct
to specify that these are all constraints the
*owner* of the set of constraints must *never
violate*.
This is done to make it easier to check that
a particular node is not violating any
constraint for a gien update, as before it
could violate constraints found both in its
local and the remote contraints.
In this commit, we make an API change that’s meant to reduce the amount
of garbage we generate when doing pathfinding or syncing nodes with our
latest graph state. Before this commit, we would always have to fully
decode the public key and signatures when reading a edge or vertex
struct. For the edges, we may need several EC operations to fully
decode all the pubkeys. This has been seen to generate a ton of
garbage, as well as slow down path finding a good bit.
To remedy this, we’ll now only ever read the *raw* bytes from disk. In
the event that we actually need to verify a signature (or w/e), only
*then* will we fully decode everything.
This commit adds Tor support. Users can set the --TorSocks flag
to specify which port Tor's SOCKS5 proxy is listening on so that
lnd can connect to it. When this flag is set, ALL traffic gets
routed over Tor including DNS traffic. Special functions for
DNS lookups were added, and since Tor doesn't natively support
SRV requests, the proxySRV function routes connects us to
a DNS server via Tor and SRV requests can be issued directly
to the DNS server.
Co-authored-by: MeshCollider <dobsonsa68@gmail.com>
Before this commit, we’d unnecessarily use a write transaction within
the FetchChannelEdgesByOutpoint. This is wasteful as the function only
actually reads items from the database, and doesn’t attempt any
mutations at all.
Fixes#481.
Prior to this commit, payments stored in the channel DB only kept a
record of the payment hash. This is a problem as the preimage is what
serves as proof of payment and a user should be able to look up this
value in the future (not just immediately after payment).
Instead of storing both the payment hash and the preimage, we store the
preimage only since the hash can be derrived from this using a SHA256.
In the RPC listpayments command, we now give the preimage in addition to
the payment hash.
This commit adds a FetchClosedChannel method to the
channeldb, which allows querying based on a known
channel point. This will be used in the nursery to
load channel close summaries, which can be used to
provide more accurate height hints when recovering
from failures.
In this commit, we add the WitnessCache sub-storage system of the
greater database. The WitnessCache is a persistent cache of all
witnesses we’ve encountered on the network. We’ll use this cache to
share any on-chain discoveries between active channels. Eventually
we’ll also use this to enforce the variant that a preimage is only to
be used ONCE on the network.
This commit adds the ChannelFlags field, of type
lnwire.FundingFlags, to the OpenChannel struct,
including serialization for database storage.
This is done to preserve the flags that were
sent during channel opening, currently used
to determine whether a channel should be made
public or not after opening.
In this commit, we fix an existing bug that arose due to incorrectly
crafting the key we use to store channel commitments. Before this
commit, we tried to copy to a slice that hadn’t been allocated yet. As
a result, the key would only have the 0x00 or 0x01 as its value. We fix
this by properly crafting the key using the built-in append function.
In this commit, we fix an existing bug wherein if we closed two
channels, then we were unable to read the channel state afterwards as
we deleted the enclosing bucket.
In this commit, we fix an existing bug wherein we failed to update the
channels state once we accepted a new commitment. As a result, after a
state transition, if the channel state was read from disk, values like
TotalMSatSent wouldn’t be properly updated.
In this commit we’ve extended the TestChannelStateTransition method to
exercise the new state transition related messages. This includes
ensuring that when we add a new dangling commitment, and then the
remote party revokes it, then the on-disk state is update accordingly.
In this commit, we update the CloseChannel method to respect the new
on-disk bucket based structure. Additionally, we now ensure that we
delete the new chainBucket.
In this commit, in addition to the renaming we’ve modified the behavior
of AdvanceCommitChainTail as follows: this method now will simply
atomically advance the commitment tail, set the new commitment to the
prior dangling commitment, and update the on-disk revocation log.
The macho expects the new revocation state to already be stored within
the channel. This method is to be called once the remote party revokes
their current commitment state.
In this commit, we add a new method: RemoteCommitChainTip. This method
allows callers to poll the database state to check if we have an
un-acked commitment for the remote party. If so, then it should be
retransmitted once a communication channel has been re-established with
the channel peer. This method will return ErrNoPendingCommit if we
don’t currently have a dangling commitment.
In this commit, we add a new method AppendRemoteCommitChain. This
method is meant to be used once we extend a new state to the remote
party, but before we actually transmit the CommitSig message. With this
method, we store a fully valid CommitDiff on disk which can be used in
the case that we need to retransmit the state to the party as they
didn’t fully receive it.
In this commit we finalized the structure of the CommitDiff struct by
adding a set of LogUpdates, and also a valid CommitSig message.
The LogUpdate struct houses a messages that were transmitted and
locked-in with the new commitment state. We include the LogIndex along
with the wire messages itself in order to be able to properly
reconstruct the update log upon restart.
The CommitSig message included should be transmitted after the set of
LogUpdates, and fully covers the new commitment state and any new (or
already present) HTLC’s that were included as part of the state.
In this commit, we modify the UpdateCommitment method to accept a full
ChannelCommitment rather than a new transaction, the sig, and a
ChannelDelta. This new structure of this method also takes advantage of
the new bucket structure of the storage schema. Additionally, this
method will now atomically swap in the new passed commitment to point
to the LocalCommitment value within the struct.
In this commit we add a new MarkAsOpen method to the OpenChannel
struct. This method replaces the existing MarkChannelAsOpen method
which targeted the database struct itself.
In this commit we comptely overhaul the existing storage of the
OpenChannel struct to use the new common serialization defined within
the codec.go file. Additionally, we’ve modified the structure of the
channel database on disk. Rather then use the existing prefix based
segmentation, everything is now bucket based. This has resulted in much
simpler and easier to follow code. The bucket progression is:
openChannelBucket -> nodeBucket -> chainBucket -> channelBucket. We add
a chainBucket as it’s possible that in the future we may have several
channels on distinct chains with a given node.
With the above changes, we’re able to delete much of the existing code
within the file, drastically reducing its size.
By adding these two fields, it is now possible to fully reconstruct the
channel’s update log from the set of HTLC’s stored on disk, as we now
properly note both the log index and HTLC index. Prior to this commit
we would simply start the new log index based on the amount of HTLC’s
that were present in the prior state.
In this commit, we restructure the OpenChannel struct to used two
distinct ChannelCommitments: one for the remote party, and one for the
local party. With this, we now partition the local state and the remote
state. Previously, we weren’t properly tracking the state of the remote
party. As a result, there were certain classes of failures we were
unable to properly recover from. With this separation, we can now
handle them, and the OpenChannel struct is slimmer and more
understandable.
In this commit we’ve added a new struct to the package:
ChannelCommitment. This sturct houses all the common data the comprises
a particular commitment state. This will soon replace the open fielded
commitmetn fields within the OpenChannel struct.
In this commit, we add a new files to the channeldb package: codec.go.
This file is similar to the way that serialization is done within the
lnwire package. The goal of this file is to reduce all the duplication
within the common serialization methods across the package.
In this commit htlc channeldb representation have been augmented
with onion blob field, and (de)serialisaion functions have been changed
to make the onion blob persistant.
This commit adds the method DisconnectBlockAtHeight to the channel
graph database, making it possible to "rewind" the database in case
a block is disconnected from the main chain. To accomplish this,
a prune log is introduced, making it possible to keep track of the
point in time where the database was pruned. This is necessary for
the case where lnd might wake up on a stale branch, and must "walk
backwards" on this branch after it finds a common block fro the
graph database and the new main chain.
This commit modifies the recently modified logic for self-channel
retransmission to exclude pruning *our* channels which haven’t been
updated since the broadcastInterval. Instead, we only re-broadcast
channels of ours that haven’t been updated in 24 hours.
Use binary.Read/Write in functions to serialize and deserialize
channel close summary and HTLC boolean data, as well as in
methods to put and fetch channel funding info. Remove lnd
implementations of readBool and writeBool as they are no
longer needed. Also fix a few minor typos.
This modifies the tests that deal serializing the Invoice type to limit
the creation date to seconds since Go1.9 added the concept of a
monotonic component to times which does not round trip through
MarshalBinary and UnmarshalBinary and therefore causes the tests to fail.
In particular, it modifies the creation dates in the randInvoice,
makeFakePayment, makeRandomFakePayment, and TestInvoiceWorkflow
functions.
This results in allowing TestOutgoingPaymentSerialization,
TestOutgoingPaymentWorkflow, and TestInvoiceWorkflow to pass.
This commit modifies the node.ForEachChannel to *also* return the
incoming edge as well as the outgoing edge. We make this modification
as when we’re doing path finding, we need to return the incoming edge
as well, since we need to use the to properly compute the time lock and
fees for transit on that edge.
Persists the state of a channel opening process after funding
transaction is confirmed. This tracks the messages sent to
the peer such that the process can be continued in case of a
restart. Also introduces that the receiver side forgets about
channel if funding transaction is not confirmed in 48hrs.
Adds a HaveNodeAnnouncement field to the LightningNode
struct, which is used to indicate if we have gotten
all the necessary information to fill the remaining
fields in the struct. If we haven't gotten a node
announcement for this specific node, then we only
know the pubkey, and can only fill that field in
the struct. Still, we should be able to add it to the
channel graph and use it for routes, as long as we
know about channels to this node.
This commit removes the RevocationDelay field from the HTLC struct as
with the latest commitment transaction scheme, it is no longer needed.
This is due to the fact the the delay is now observed when an on-chain
HTLC claim is attempted, rather than from Shane the HTLC itself has
been broadcast.
This commit adds a new method to the OpenChannel struct:
InsertNextRevocation. This functionality is required in order to
properly implement the new funding workflow and commitment transaction
state machine.
In the prior state machine, we utilized the “initial revocation window”
which would provide both sides with the necessary revocation keys that
they needed to create new commitment states. The version of the
commitment state machine within the specification has been simplified a
bit, and instead essentially incorporates a revocation window of 1. The
window is initially filled at the _opening_ of the commitment
transaction, then never extended, only “slid down” from there on.
This commit modifies the HTLC struct within the channeldb struct in
order to accommodate the additional storage that’s now required as a
result of the change to 2-of-2 HTLC covenant transactions within the
commitment transaction.
This commit is the next phase in a rather major overhaul of the
OpenChannel struct. With this commit we remove the old ours vs theirs
semantics with channel related state variables, and instead encapsulate
all local vs remote state into a ChannelConfig structure for each party
within the channel.
This commit introduces two new structs to the channeled package:
ChannelConfig and ChannelConstraints.
The ChannelConfig struct will eliminate many of the fields on the
OpenChannel struct (ours vs theirs), in favor of having a single config
for both sides (local and remote).
The ChannelConstraints struct will house the channel flow control and
HTLC policies for both sides. The fields of this struct will be used to
control the flow of HTLC’s within the channel. In the occasion that any
of these constraints are violated, either the connection, or the
channel entirely is to be failed.
This commit removes a database migration that was added to fix a bug
introduced within the first alpha release. This migration is no longer
needed as the underlying database scheme has changed since then.
This commit adds the possibility for the initiator of a
channel to send the update_fee message, as specified
in BOLT#2. After the message is sent and both parties
have committed to the updated fee, all new commitment
messages in the channel will use the specified fee.
The btclog package has been changed to defining its own logging
interface (rather than seelog's) and provides a default implementation
for callers to use.
There are two primary advantages to the new logger implementation.
First, all log messages are created before the call returns. Compared
to seelog, this prevents data races when mutable variables are logged.
Second, the new logger does not implement any kind of artifical rate
limiting (what seelog refers to as "adaptive logging"). Log messages
are outputted as soon as possible and the application will appear to
perform much better when watching standard output.
Because log rotation is not a feature of the btclog logging
implementation, it is handled by the main package by importing a file
rotation package that provides an io.Reader interface for creating
output to a rotating file output. The rotator has been configured
with the same defaults that btcd previously used in the seelog config
(10MB file limits with maximum of 3 rolls) but now compresses newly
created roll files. Due to the high compressibility of log text, the
compressed files typically reduce to around 15-30% of the original
10MB file.
This commit modifies the OpenChannel struct to include the full short
channel ID rather than simply the opening height. This new field will
be needed by an upcoming change to uniformly switch to using short
channel ID’s when forwarding HTLC’s due to the change in per-hop
payloads.
This commit expands the field within the OpenChannel struct in order to
start tracking the height that the funding transaction was initially
broadcast. Other sub-systems within lnd can now use this data to give a
more accurate height hint to the ChainNotifier, or to use during the
funding workflow to decide if a channel should be forgotten after it
fails to confirm for N blocks.
This commit modifies the name of a field in the OpenChannel struct to
better reflect its actual usage within this protocol. The FeePerKw
represents the amount of satoshi to be paid as fees per kilo-weight.
This field is set at the opening of a transaction and will be able to
be updated properly via the usage of the update_fee method.
In order to cleanly handle shutdowns and restarts during state machine operation, the fee for the current
commitment transaction must be persisted. This allows the fee to be
reapplied when the current state is reloaded.
This commit adds the total time locked balance of a closed channel at
closure time to the CloseChannelSummary struct. With this information,
we now provide sub-systems within the daemon further knowledge which
can aide them in properly handling querying for the state of pending
close transactions, or if they should watch a channel for closure or
not.
This commit modifies the OpenChannel structure on-disk to also track
that opening height of a channel. This change is being made in order to
make and more light client friendly. A follow up commit will modify
several areas of the codebase to use this new functionality.
This commit introduces a new method to the ChannelGraph struct:
ChannelView. This struct returns all the outpoints that represent the
set of active channels within the network. The set of items returned by
this new method will possibly shrink with each call to `PruneGraph`,
and possibly be expanded by each call to `AddChannelEdge`.
The graph pruning tests have been updated to ensure the description
above holds true.
This commit adds a new method to the database which allows callers to
mark a channel as fully closed once certain conditions have been
reached. If a channel was cooperatively closed, then it can be marked
as fully closed as soon as the channel has been confirmed. If the
channel was marked as force closed, then it should be marked as closed
as soon as _all_ the funds in limbo have been swept.
This commit builds upon the prior commit by adding a new method to
allows callers to query the state of all closed (fully) and pending
closed channel within the database.
This commit expands the close summaries within the database to include
additional information such as: how the channel was closed, if the
channel if fully closed, and other balance information. This new
information will be used in an upcoming expansion of the
pendingChannels RPC to display more comprehensive data w.r.t where the
node’s funds are and how soon they’ll be fully available.
The data stored in the closeChannel bucket has been expanded to include
the above. Additionally the CloseChannel method on the OpenChannel
struct now takes in the additional information that details how and who
it was closed by.
This commit borrows heavily from the initial work in #179 by @halseth.
This commit modifies the ForEachNode on the ChannelGraph and
ForEachChannel on the LightningNode struct to accept a database
transaction as its first argument. With this change, it’ll now be
possible to implement graph traversals that typically required a nested
loop with all the vertex loaded into memory using the callback API
instead:
c.ForEachNode(nil, func(tx, node) {
node.ForEachChannel(tx, func(…) {
})
})
This commit implements an easy optimization by using bolt db’s Batch
method when writing payment details to disk. The AddPaymnent method can
be concurrently called by thousands of grouting due to the way the
payment dispatch pipeline is architected. With this commit, we shave of
a significant amount of running time when users are sending thousands
of payments a second as what would’ve been thousands of writes can now
be coalesced into one or two writes!
Originally we adding the edge without proof in order to able to use it
for payments path constrcution. This method will allow us to populate
the announcement proof after the exchange of the half proofs and
constrcutrion of full proof finished.
In case if the channel shouldn't be announced to the rest of the network
the proof, which is needed to announce the channel, will not be
populated, fot that reason the ability to store the empty proof has
been added.
Minor change to server.go to add ExternalIPs to
channeldb.LightningNode. Also, added a test that utilizes this
functionality and exercises multiple addresses in NodeAnnouncement.
This commit modifies address handling in the NodeAnnouncement struct,
switching from net.TCPAddr to []net.Addr. This enables more flexible
address handling with multiple types and multiple addresses for each
node. This commit addresses the first part of issue #131 .
This commit adds a new method to the channel state: RevocationLogTail.
This new method will return the information concerning the latest
revoked state of the remote party’s commitment chain.
This new data can be used to properly initialize the states of the
in-memory commitment chains on node start up.
This commit removes all instances of the fastsha256 library and
replaces it with the sha256 library in the standard library. This
change should see a number of performance improvements as the standard
library has highly optimized assembly instructions with use vectorized
instructions as the platform supports.
This commit modifies the path finding routines to properly use the new
channel edge related API exposed by the database. Additionally, a new
type `ChannelHop` has been introduced which couples an edges routing
policy with the capacity and origin chain of the channel.
This commit alters the return value of PrunedGraph to me a bit more
useful: the function now returns all the channels that were closed when
processing the block (slice of spent outpoints). With this information,
callers gain greater visibility into exactly which channels were
closed. This can be used in higher levels to present detailed summaries
of how blocks affect closed channels.
This commit splits the prior ChannelEdge struct into two distinct
structs: ChannelEdgeInfo and ChannelEdgePolicy. The info structs stores
the _full_ information that was used to advertise the channel, while
the policy struct contains the information that’s needed in order to
use the information for routing purposes.
With this split we can eliminate a number of hacks within the rest of
the codebase that were added as a result of data unavailability if one
or neither edge was present.
Finally a bit of field renaming has taken place (Exipiry ->
TimeLockDelta), etc.
When a pending channel is persisted and then reloaded upon system startup
it's necessary to also persist the number of confirmations that will be required
before the pending channel can be opened.
In order to facilitate persistence during the funding process, added
the isPending flag to channels so that when the daemon restarts, we can
properly re-initialize the chain notifier and update the state of
channels that were going through the funding process.
In this commit the initial implementation of revocation hash
generation 'elkrem' was replaced with 'shachain' Rusty Russel
implementation which currently enshrined in the spec. This alghoritm has
the same asymptotic characteristics but has more complex scheme
to determine wish hash we can drop and what needs to be stored
in order to be able to achive full compression.
Fix SetStateNumHint and GetStateNumHint to properly
set and get the stateNumHints using the lower 24 bits
of the locktime of the commitment transaction as the
lower 24 bits of the obfuscated state number and the
lower 24 bits of the sequence field as the higher 24
bits.
This commit fixes a bug that was introduced when we moved to using
64-bit integers for storing the revocation log state. When we made this
change, we forgot to increase the size of the buffer which stores the
key for the particular channel state from 40 to 44 bytes to account for
the 4 additional bytes in the new 64-bit integer.
This bug has been fixed by properly sizing the key buffer. We’ve also
added an additional test to ensure that we retrieve the proper state
after multiple state updates.
This commit modifies the running update count within all ChannelDelta’s
to track the number of updates using a uint64 rather than a uint32.
This change reflects the fact that the obsfucated commitment hints are
to be encoded using a 48-bit integer, rather than a 32-bit integer.
This commit fixes a bug which would previously lead to corruption of
the channel state when a node had one or more channels open and one of
them was closed either forcibly or cooperatively. The source of the bug
itself as a typo: rather than using the construed `deliveryKey`
variable to fetch/put/delete the delivery scripts, `deliveryScriptsKey`
(the key prefix itself) as used. This bug would cause the database to
be unable to read _any_ channel from the database after one was
deleted, as each channel would actually be reading/writing-to the
_exact same_ delivery script.
The fix for the bug itself is simple: eliminate the typo.
This commit addresses some lingering TODO’s which ensure that related
state to a channel is properly deleted by the CloseChannel method.
Previously the values for the respective dust-limits of either side,
the on-disk HTLC’s, and any entries the revocation log for the channel
weren’t being properly deleted.
Additionally, we now modify the checks within the unit tests to ensure
that we can still read the channel from disk w/o running into an error
(thought the slice will be blank), and also the the revocation log is
properly garbage collected.
This commit fixes a panic that would arise when the daemon attempts to
query for a channel that doesn’t currently exist. The bug was the
result of a typo which checked for the nil existence of the incorrect
variable.
This commit fixes a prior bug in the graph database due to an invalid
assumption that both channel edges would _always_ be advertised. This
assumption is invalid, as it’s up to a node’s policy if the advertise
their direction of the channel.
The fix for this assumption is straight forward: ErrEdgeNotFound is no
longer a critical error, instead a nil pointer will now be passed into
the passed callback function.
This commit makes a large number of minor changes concerning API usage
within the deamon to match the latest version on the upstream btcsuite
libraries.
The major changes are the switch from wire.ShaHash to chainhash.Hash,
and that wire.NewMsgTx() now takes a paramter indicating the version of
the transaction to be created.
This commit modifies the error propagating behavior within the
HasChannelEdge struct. Rather than exiting the function early when a
single edge isn’t found, we instead continue to also possibly retrieve
the second directional edge.
With this change, we avoid a potential infinite gossiping loop in the
routing package that would result if we’d seen one edge but not the
other. In this case the timestamps returned for *both* edges would
always be zero, causing the daemon to always accept and rebroadcast the
announcement putting all other connected lnd’s into the same loop.
This commit modifies the new payment module within the database to
match the coding style of the rest of the package and the project as a
hole. Additionally, a few fields have been renamed, and the extra
timestamp added to the OutgoingPayment struct has been removed as
there’s already a CreationTime field within the Invoice struct that’s
embedded within the OutgoingPayment struct.
This commit modifies the FetchChannelEdgesByID slightly to use a
read-only transaction rather than a write-only transaction. As a result
we’ll no longer extraneously consume a writer’s slot when we’re only
reading data from the database.
This commit modifies the HasChannelEdge function to _always_ return
true if we know of the channel edge, meaning that it was previously
added on announce.
This change fixes a minor bug present in the code which would result in
extraneous re-transmissions of updates received by the new routing
package.
This commit adds a utility method whcih utilizes the edge index bucket
and allows caller to look up the channel ID of a channel by its funding
outpoint. This can be used to populate RPC’s with additional
information and also to provide users with an additional query
interface to build channel explorers, etc.
Previously, the edge index bucket which maps a channelPoint ->
channelID wasn’t properly created one start up during the initial
creation of the database. This caused some extraneous failure as
queries would unnecessarily fail with bucket non-existence errors.
To fix this we now properly create the bucket on start up if the
database doesn’t exist, and also properly delete the bucket within the
Wipe() function.
This commit fixes a minor bug in the ForEachChannel method of the
ChannelGraph struct. Rather than ErrGraphNoEdgesFound being returned if
either of the edge related buckets hadn’t been created yet,
ErrGraphNodesNotFound was being returned.
To fix this bug, we now properly return ErrGraphNoEdgesFound.
Additionally a mental note to roasbeef has been left as the current
code currently assumes that eventually both directions of the channel
edge will be advertised. However, this may not necessarily be the case
in a live network, since a side chooses to preferentially advertise a
channel or not.
This commit adds to new functions to the ChannelGraph struct which
allow the callers to query for the existence or non-existence of a
vertex (node) or edge (channel) within the graph. In addition to
returning whether the edge exists, the functions will also return the
last time the state has been modified for the edge or vertex. This will
allow callers to ensure that only the most up to date state is
committed to disk.
This commit adds support for channel graph pruning, which is the method
used to keep the channel graph in sync with the current UTXO state. As
the channel graph is essentially simply a subset of the UTXO set, by
evaluating the channel graph with the set of outfits spent within a
block, then we’re able to prune channels that’ve been closed by
spending their funding outpoint. A new method `PruneGraph` has been
provided which implements the described functionality.
Upon start up any upper routing layers should sync forward in the chain
pruning the channel graph with each newly found block. In order to
facilitate such channel graph reconciliation a new method `PruneTip`
has been added which allows callers to query current pruning state of
the channel graph.
This commit adds an additional check within CloseChannel to ensure that
sub-systems attempting to delete the channel one after the other (in
the event of any sort of closure) doesn’t result in an extraneous
error.
To fix this, we now check if the channel exists before attempting a
deletion. If the channel doesn’t exist, then we simply exit early with
a nil error.
This commit modifies the LightningNode.ForEachChannel method to give
the caller the option of re-using an existing database transaction
instead of always creating a new db transaction with each invocation.
Internally boltdb will run into an error/dead-lock if a nested
transaction is attempted.
Such an action might be attempted if one were to use the traversal
functions in a path finding algorithm. Therefore in order to avoid
that after, we now allow the re-use of transactions to facilitate
nested calls to ForEachChannel.
Go-fmt files. Refactored code according to the guidelines.
Enhanced payment test: add error checking
and individual context for each API call.
Add Timestamp field to payment struct.
This commit introduces a new capability to the database: storage of an
on-disk directed channel graph. The on-disk representation of the graph
within boltdb is essentially a modified adjacency list which separates
the storage of the edge’s existence and the storage of the edge
information itself.
The new objects provided within he ChannelGraph carry an API which
facilitates easy graph traversal via their ForEach* methods. As a
result, path finding algorithms will be able to be expressed in a
natural way using the range methods as a for-range language extension
within Go.
Additionally caching will likely be added either at this layer or the
layer above (the RoutingManager) in order keep queries and outgoing
payments speedy. In a future commit a new set of RPC’s to query the
state of a particular edge or node will also be added.
This commit adds a new method to the `OpenChannel` struct:
CommitmentHeight(). This method allows multiple callers holding the
same instance of an OpenChannel struct tied to the same on-disk channel
to consistently query the current commitment height for a channel. Such
a modification will prove useful later as sections of the code-base are
separated in order to allow more vigilant watching of channel breaches.
This commit performs a slight refactoring of the internals (and API) of
the [Fetch|Put]Meta methods. The changes are rather minor and simply
eliminate the conditional branching structure with usage of an internal
function. This new form is much easier to follow.
This commit modifies the composition of the boltdb pointer within the
DB struct to use embedding.
The rationale for this change is that the daemon may soon store some
semi-transient items within the database which requires us to expose
the boltdb’s transaction API. The logic for serialization of this data
will likely lie outside of the channeldb package as the items that may
be stored in the future will be specific to the current sub-systems
within the daemon and not generic channel related data.
This commit unexports the SyncVerions PR, in favor of making it private
and moving it into the .Open() method. With this change, callers no
longer need worry about ensuring the database version is up to sync
before usage, as it will happen automatically once the database is
opened.
This commit also unexports some additional variables within the package
that don’t need be consumed by the public, and exports the
DbVersionNumber attribute on the meta struct. Finally some minor
formatting fixes have neen carried out to ensure the new code more
closely matches the style of the rest of the codebase.
In this commit the upgrade mechanism for database was added which makes he current schema rigid and upgradeable. Additional bucket 'metaBucket' was added which stores
key that house meta-data related to the current/version state of the database. 'createChannelDB' was modified to create this new bucket+key during initializing. Also
backup logic was added which makes a complete copy of the current database during migration process and restore the previous version of database if migration failed.
This commit adds persistence for a channel’s
TotalSatoshis[Sent|Received] fields. Also, the functions that perform
this persistence were renamed from [put|delete|fetch]ChanTotalFlow to
[put|delete|fetch]ChanAmountsTransferred.
This commit modifies the ChannelType enum to no longer use iota as
changes in the definition would cause the values to shift, breaking the
long-term stability required for persistence. Instead, we now select
values manually to indicate the particular channel type.
This commit modifies the existing OpenChannel struct slightly to store
a bool which indicates if we were the one who initially initiated the
channel or not. This information is relevant as in the current draft of
the specification, much of the fee related negotiation is contingent on
who initiated the channel.
This commit removes all prior fee tracking attributes along with the
persistence code from OpenChannel. The rationale is that fees actually
don’t exist at the channel level, and instead should be tracked at
higher level of abstraction as fees come from the inbound/outbound
satoshi spread.
This commit modifies the OpenChannel struct along with all related
persistent to additional store a single byte which indicates the exact
type of the channel. This may be useful in the future as higher level
behavior may change depending on the precise type of the channel.
This commit introduces a new method to channeldb: ‘FetchAllChannels’.
This method can be used to obtain the state of all active (currently
open) channels within the database. This method can be used for compute
basic channel-based metrics or exposed as an RPC in order to allow
clients to display/query channel data.
This commit adds a new method ‘FullSyncWithAddr’ which is identical to
the existing ‘FullSync’ method other than it also creates an
association from the channel to a LinkNode object within the database.
This new method is required in order to create persistent links between
channels and link nodes which will later allow the development of
heuristics which decided when it “makes sense” to close a channel due
to inactivity. Additionally, this new association will allow for a
sub-system within the daemon to attempt to establish persistent
connections out to all LinkNodes in order to strive for channel
availability.
This commit slightly modifies the existing structure of the channeldb
scheme to replace the former concept of a “nodeID” with simply the
compressed public key of the remote node. This change paves the way for
adding useful indexes mapping a node to all it’s active channels and
the other way around.
Additionally, the current channeldb code was written before it was
agreed by many of those implementing Lightning that a node’s ID will
simply be its compressed public key.
This commit adds a new bucket to the database which is dedicated to
storing data pertaining to p2p related reachability for direct channel
counter parties. The data stored in this new bucket can be used within
heuristics when deciding to unilaterally close a channel due to
inactivity. Additionally, all known reachable IP addresses for a
particular LinkNode are to be stored and updated within the database in
order to facilitate the establishment of persistent connections to
direct channel counter parties.
This commit consists of a mass variable renaming to call the pkScript being executed for segwit outputs the `witnessScript` instead of `redeemScript`. The latter naming convention is generally considered to be reserved for the context of BIP 16 execution. With segwit to be deployed soon, we should be using the correct terminology uniformly through the codebase.
In addition some minor typos throughout the codebase has been fixed.
This commit modifies the on-disk storage of invoices to stop the
optional fields (memo+receipt) on-disk as variable length byte arrays.
This change saves space as the optional fields now only take up as much
space as is strictly needed, rather than always being padded out to max
size (1KB).
This commit adds a new invoice related method: FetchAllInvoices. This
method allows callers to query the state of all invoices currently
stored within the database. The method takes a toggle bit which
determines if only pending (unsettled) invoices should be returned, or
if they al should be.
This commit moves the location of the invoice counter key which is an
auto-incrementing primary key for all invoices. Rather than storing the counter
in the same top-level invoice bucket, the counter is now stored within the
invoiceIndex bucket. With this change, the top-level bucket can now cleanly be
scanned in a sequential manner to retrieve all invoices.
This commit adds the necessary database functionality required for a
high-level payment invoice workflow. Invoices can be added dealing the
requirements for fulfillment, looked by payment hash, and the finally
also settled by payment hash. For record keeping and the possibility of
reconciling future disputes, invoices are currently never deleted from
disk. Instead when an invoice is settled a bit is toggled indicating as
much.
The current invoiceManger within the daemon will be modified to use
this persistent invoice store, only storing certain “debug” invoices in
memory as dictated by a command line flag.
This commit introduces the concept of a manually initiated “force”
closer within the channel state machine. A force closure is a closure
initiated by a local subsystem which broadcasts the current commitment
state directly on-chain rather than attempting to cooperatively
negotiate a closure with the remote party.
A force closure returns a ForceCloseSummary which includes all the
details required for claiming all rightfully owned outputs within the
broadcast commitment transaction.
Additionally two new publicly exported channels are introduced, one
which is closed due a locally initiated force closure, and the other
which is closed once we detect that the remote party has executed a
unilateral closure by broadcasting their version of the commitment
transaction.
This commit splits the previously added RecordChannelDelta method into
two distinct methods: UpdateCommitment and AppendToRevocationLog. The
former method is to be used once the local party revokes their current
commitment, and the latter method to be used once the remote party
revokes their current commitment.
With the addition of the UpdateCommitment method, the active HTLC’s
from the local node’s point of view are now persisted to disk.
Snapshots returned by the channel now also includes the current set of
active HTLC’s. In order to maintain thread safety the channels mutex is
now grabbed within methods which modify/read state but don’t do so
solely via a boltDB transaction.
The tests have been updated to account for the storage of HTLC’s needed
in order to assert proper behavior.
This commit removes the revocation hash/keys from the channel deltas.
In the case of an uncooperative closure, we can efficiently re-generate
the proper elkrem pre-image so this storage was completely unnecessary
This commit implements a state update log which is intended the record
the relevant information for each state transition on disk. For each
state transition a delta should be written recording the new state. A
new method is also provided which is able to retrieve a previous
channel state based on a state update #.
At the moment no measures has been taken to optimize the space
utilization of each update on disk. There are several low-hanging
fruits which can be addressed at a later point. Ultimately the update
log itself should be implemented with an append-only flat file at the
storage level. In any case, the high level abstraction should be able
to maintained independent of differences in the on-disk format itself.
This commit removes the storage+encryption of private keys within
channeldb. We no longer need to encrypt these secrets before storing as
the base wallet is now expected to retain full control of these secrets
rather than the database.
As a result, we now only store public keys within the database.
This commit removes the EncryptorDecryptor interface, and all related
usage within channeldb. This interface is no longer needed as wallet
specific secrets such as private keys are no longer stored within the
database.
This commit changes the current behavior around channeldb.Wipe().
Previously if a channel had never been closed, and a wipe was
attempted, then wipe operation would fail and the transaction would be
rolled back.
This commit fixes this behavior by checking for bolt.ErrBucketNotFound
error, and handling the specific error as a noop.
This commit fixes a bug caused by overriding the prefix key for storing
commitment keys with the first few bytes of a channel’s channel point.
Once a channel was deleted, then all future channels would result in a
panic due to a nil pointer deference since the prefix key was mutated,
causing all future stores/gets to fail.
We now also store their current revocation hash which is given to us
along with the revocation key once an initial HTLC is added to a
commitment transaction.
This commit updates the stored on-disk channel state to store a current
revocation key rather than a revocation hash. This change coincides
with the new commitment transaction format which uses revocation keys
rather than hashes.
Additionally, this commit updates the decoding/encoding of local+remote
elkrem trees to the latest changes in the elkrem API.
This commit fixes a bug which would potentially cause a panic if a
channel returned from FetchOpenChannels attempted to access the
internal pointer to the database.
To fix this bug, the pointer is now properly set once the channel has
been loaded from the database.
This commit introduces the concept of “closing” an already active
channel. Closing a channel causes all the channel state to be purged
from the database, and also triggers the creation of a small “summary”
kept concerning details of the previously open channel.
This commit also updates the previous test case(s), and includes the
close channel bucket in the database deletion in the .Wipe() method.
This commit overhauls the current schema for storing active channels in
order to support tracking+updating multiple open channels for a
particular peer.
Channels are now uniquely identified by an output (txid:index) rather
than an arbitrary hash value. As a result, the funding transaction is
no longer stored, as only the txin is required to lookup the original
transaction, and to sign for new commitment states.
A new bucket, nested within the bucket for a node’s Lightning ID has
been created. This new bucket acts as an index to the active channels
for a particular peer by storing all the active channel points as keys
within the bucket. This bucket can then be scanned in a linear fashion,
or queried randomly in order to retrieve channel information.
The split between top-level, and channel-level keys remains the same.
The primary modification comes in using the channel ID (the funding
outpoint) as the key suffix for all top-level and channel-level keys.
The state of OpenChannel on disk has now been partitioned into several
buckets+keys within the db. At the top level, a set of prefixed keys
stored common data updated frequently (with every channel update).
These fields are stored at the top level in order to facilities prefix
scans, and to avoid read/write amplification due to
serialization/deserialization with each read/write.
Within the active channel bucket, a nested bucket keyed on the node’s
ID stores the remainder of the channel.
Additionally OpenChannel now uses elkrem rather than shachain, delivery
scripts instead of addresses, stores the total net fees, and splits the
csv delay into the remote vs local node’s.
Several TODO’s have been left lingering, to be visited in the near
future.
Commit includes basic tests for Open/Create. Additionally, rather than
relying on btcwallet’s addmgr for encryption/decryption, this package
now exposes a simple crypto system interface.
* Initial draft of brain dump of chandler. Nothing yet set in stone.
* Will most likely move the storage of all structs to a more “column”
oriented approach. Such that, small updates like incrementing the total
satoshi sent don’t result in the entire struct being serialized and
written.
* Some skeleton structs for other possible data we might want to store
are also included.
* Seem valuable to record as much data as possible for record keeping,
visualization, debugging, etc. Will need to set up a time+space+dirty
cache to ensure performance isn’t impacted too much.
