It was being considered a misbehaviour from the intermediate hop which
penalized that hop, which is a bit harsh considering mobile nodes. We
change it be considered as other channel update types and only penalize
the channel that failed.
Having it set to nil caused https://github.com/lightningnetwork/lnd/issues/5115
The problem was several layers removed from the fix. The link decides to
clean up a `fwdPkg` only if it's completed, otherwise it renotifies the
HTLCs. A package is only set to complete if it's `addAck` and
`settleFail` filters are full. For forwarded HTLCs, the `addAck` was
never being set so it would never be considered complete under this
criteria.
`addAck` is set for an HTLC when signing the next commitment TX in the
`LightningChannel`. The path for this is:
* `LightningChannel#SettleHtlc` adds the HTLC to `localUpdates`
* `LightningChannel#SignNextCommitment` builds the `ackAddRef` for all
updates with `SourceRef != nil`.
* `LightningChannel#SignNextCommitment` then passes the list of
`ackAddRef` to `OpenChannel#AppendRemoteCommitChain` to persist the new
acks in the filter
Since `SourceRef` was nil for interceptor packages, `SignNextCommitment`
ignored it and the ack was never persisted.
distance map now holds the edge the current path is coming from,
removing the need for next map.
Both distance map and distanceHeap now hold pointers instead of the full
struct to reduce allocations and copies.
Both these changes reduced path finding time by ~5% and memory usage by
~2mb.
Pre-sizing these structures avoids a lot of map resizing, which causes
copies and rehashing of entries. We mostly know that the map won't
exceed that size, and it doesn't affect memory usage in any significant
way.
Calling `ForEachNode` hits the DB, and allocates and parses every node
in the graph. Walking the channels also loads nodes from the DB, so this
meant that each node was read/parsed/allocated several times per run.
This reduces runtime by ~10ms and memory usage by ~4mb.
The server was kept alive long after it stopped being used. This caused
problems for services using long-lived GRPC connections which might be
created before wallet unlocked. They got stuck connected to the wallet
unlock service needing a restart.
Juan Pablo Civile