limbo balance is stuck on a waiting to close channel

[ubenmackin]

Background

Power went out, and client didn't gracefully shutdown. Now, many weeks later, I have a channel balance that shows "limbo":

{
"total_limbo_balance": "82973",
"pending_open_channels": [],
"pending_closing_channels": [],
"pending_force_closing_channels": [],
"waiting_close_channels": [
{
"channel": {
"remote_node_pub": "026b105ac13212c48714c6be9b11577a9ce10f10e1c88a45ce217e6331209faf8b",
"channel_point": "bd7ae7389298573d8e092fed7d1d5040536d8aa7fc6d679d5e381e99fd11486f:1",
"capacity": "83154",
"local_balance": "82973",
"remote_balance": "0"
},
"limbo_balance": "82973"
}
]
}

Your environment

• version of lnd
  0.4.2-beta commit=7cf5ebe2650b67...1f1d6e19
• which operating system (uname -a on *Nix)
  Linux bitnode 4.15.0-23-generic Figure out packaging for language-specific gRPC libs #25-Ubuntu SMP Wed May 23 18:02:16 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
• version of btcd, bitcoind, or other backend
  bitcoins v0.16.1
• any other relevant environment details

Steps to reproduce

Unsure exactly what caused the issue.

Expected behaviour

Balance shouldn't be stuck in limbo, or there should be a way to close out the channel.

Actual behaviour

Channel seems stuck, and balance it stuck as well.

[Roasbeef]

The channel you linked was closed back in May, I'm guessing you created it with a version of lnd before 0.4.1? If so, then you may have this this bug, which has subsequently been fixed. As a result, new versions of lnd won't be able to get into this configuration.

If you send us some start up logs, then we'll be able to confirm this.

[ubenmackin]

Ok, will do as soon as I get home.

[ubenmackin]

@Roasbeef I started, let it run for a few minutes, then stopped. Here is the log output. If you need more, let me know.

Link to Log

Edit: You are correct that I had originally done it under 0.4 (I believe). I have updated my LND twice, once to 0.4.1 and then more recently to 0.4.2.

This was all done under autopilot, I haven't actually run any manual channel commands yet.

[Roasbeef]

2018-06-25 18:14:13.435 [ERR] CNCT: Remote node broadcast state #37, which is more than 1 beyond best known state #34!!!

Looks like it's suffering from bugs in 0.4, that were later fixed in 0.4.2+. It's even the case that the channel was out of sync by several states, which is something that should never happen (and won't in the current master). As a result, the output will remain on the chain unless your node knew of the revocation point for that state (which was ahead of where it was). My current guess is that the state was replicated, and as a result diverged.

[ubenmackin]

@Roasbeef Just so I am clear, the amount is lost, right? Not a huge deal (it wasn't a large amount), but I just want to be clear.

Also, is there any way to clear it out of my client, or will it just always be there in a limbo state?