Lnd doesn't currently do a graceful shutdown when a `SIGTERM` is issued.
To ensure a graceful shutdown, call `stopDaemon` on the gRPC interface
prior to terminating lnd. If a graceful shutdown is unsuccessful then
force kill the daemon.
Make sure that if a user of ours is using neutrino that we flag their
channels as private when they manually open them. Reason for this is
because light consumer clients are not expected to have high uptime.
However we cannot hardcode channels as private every time because we
do support more power users driving their remote nodes or BTCPay
Server with Zap. So we only flag channels as private if the
activeConnection is local.
During the sync process there are log message that relate to syncing
filters and filter headers. This fixes an issue where we were parsing
messages about syncing filters as if they were filter headers, which
would cause the progress indicator to jump around during the middle of
the sync process due to it thinking that more filter headers had been
synced than was actually the case.
Do not set `recovery_window: 250` in our call to `initWallet` when
creating a new wallet, as this causes lnd to scan the entire chain for
non-existent funds which slows the lnd sync process.
See https://github.com/lightningnetwork/lnd/issues/1744
Jack Mallers
Tom Kirkpatrick