Unfortunately, we can't use the graph cache directly here as it doesn't
include all the information we need, since it only includes the minimal
amount of information needed for path finding.
Instead, we use a simple mutex guarded cache that evicts itself after a
certain interval. The default is set small enough that most users
shouldn't really notice. We also provide a way to disable the cache, and
also disable the cache in the itests.
Fixes https://github.com/lightningnetwork/lnd/issues/1232
The latest version of btcd allows its stall handler to be disabled. We
use that new config option to make sure the mining btcd node and the lnd
chain backend btcd node aren't disconnected if some test takes too long
and no new p2p messages are exchanged.
Fixes#4383 by adding a new SafeCopyMacaroon function that correctly
clones all caveats and prevents modifications on the copy from affecting
the original.
From the error in the itest output log it is not clear whether scraping
the profile page caused a test to fail or whether it was just a
follow-up error. We make it a bit more clear with an added message.
In this commit, we attempt to fix a race condition that may occur in the
current AMP and MPP tests.
It appears the following scenario is possible:
* The `mppTestContext` [is used to create 6 channels back to
back](https://github.com/lightningnetwork/lnd/blob/master/lntest/itest/lnd_amp_test.go#L43)
* The method used to create the channel ends up calling
[`openChannelAndAssert`](edd4152682/lntest/itest/lnd_mpp_test.go (L300))
which'll open the channel, mine 6 blocks, [then ensure that the
channel gets
advertised](edd4152682/lntest/itest/assertions.go (L78))
* Later on, [we wait for all nodes to hear about all channels on the
network
level](https://github.com/lightningnetwork/lnd/blob/master/lntest/itest/lnd_amp_test.go#L62)
I think the issue here is that we'll potentially already have mined 30
or so blocks before getting to the final nodes, and those nodes may have
already heard about the channel already. This may then cause their
[`lightningNetworkWatcher`](edd4152682/lntest/node.go (L1213))
goroutine to not properly dispatch this, since it's assumed that the
channel hasn't been announced (or notified) when the method is called.
One solution here is to just check if the channel is already in the
node's graph or not, when we go to register the notification. If we do
this in the same state machine as the watcher, then we ensure if the
channel is already known, the client is immediately notified. One thing
that can help us debug this more in the future is adding additional
logging in some of these helper goroutines so we can more easily track
the control flow.
This commit implements this solution by checking to ensure that the
channel isn't already known in our channel graph before attempting to
wait for its notification, as we may already have missed the
notification before this registration request came in.
In #5364 we added a new error path in the `StopDaemon` method to return
an error if shutdown was attempted while a rescan/recover instance was
in progress. Since the wallet actually won't fully stop (atm)
mid-recovery, the call effectively didn't do anything in that scenario,
so we started to return an error to properly reflect that. However this
causes certain itests to fail, as during recovery, the stop attempt will
fail leading to the test itself failing.
In this commit, we wrap the calls to stop a running daemon within a
`wait.NoError` call so we'll continually try to shut down the daemon
rather than quit on the first try.
Fixes#5423.
In some rare instances it can happen that the nodes don't find each
other again after one of them has been re-created and the other one has
been restarted in the SCB tests. By making sure the re-created has the
same P2P port again as before, we make sure they can connect to each
other again successfully for executing DLP.
This commit adds a new "waiting to start" state which may be used to
query if we're still waiting to become the cluster leader. Once leader
we advance the state to "wallet not exist" or "wallet locked" given
wallet availablity.
This commit replaces most of the hard coded 10, 15, 20 and 30 second
timeouts with the default timeout. This should allow darwin users to
successfully run the parallel itests locally as well.
With the new btcd version we can specify our own listen address
generator function for any btcd nodes. This should reduce flakiness as
the previous way of getting a free port was based on just picking a
random number which lead to conflicts.
We also double the default values for connection retries and timeouts,
effectively waiting up to 4 seconds in total now.
In some tests we moved channeld.db to a temp location in order to
"time travel". This commit extends the existing semantics by moving all
files, including embedded etcd db too besides the channeld.db file.
This commit adds the icase name to the log filename, to make it simpler
to find problematic tests. Additionally after this commit we'll restart
Alice and Bob (the base harness nodes) before each icase to start with a
clean state.
To allow running multiple test tranches in parallel, we need a way to
make sure the TCP ports don't collide. We'll work with offsets for the
ports, using a different offset for each tranche.
This changes the wait during node connection to check both for the
existance as well as for the validity of the tls cert and macaroon
files.
This ensures that nodes in the process of starting up don't inadvertedly
cause a connection error due to not yet having written the entire file.
