error combining partial signature: final signature is invalid

[MegalithicBTC]

I am making this a discussion instead of an issue as I'm not sure if this is a problem with LND or it's just some kind of data corruption on my end.

The final endpoint of this error are these logs:

2024-10-02 22:36:15.945 [ERR] CNCT: ChannelArbitrator(48073b7767573658333a14956e2109f74331498ea4686a4c3340914289d37173:0): unable to advance state: unable to combine musig partial sigs: error combining partial signature: error combining partial signature: final signature is invalid
2024-10-02 22:36:15.945 [INF] CNCT: ChainArbitrator shutting down...

However, I want to describe as fully as I can how I got here.

---> HERE ARE THE COMPLETE LOGS YOU CAN DOWNLOAD:
lnd.log

This is an LND node, run in lighting-terminal integrated mode:

user@aw75:~/lightning-terminal$ lncli --version
lncli version 0.18.3-beta commit=lightning-terminal-v0.13.4-alpha

This node is running on Testnet. It was working well until a few hours ago, corresponding to line 15888 in the attached lnd.log....

2024-10-02 03:49:00.240 [TRC] LNWL: Re-establishing timed out ZMQ transaction connection
2024-10-02 03:49:00.828 [INF] GRPH: Processed channels=0 updates=1 nodes=3 in last 3m11.575116312s
2024-10-02 03:49:05.135 [DBG] HLCK: Health check: chain backend invoking failure callback
2024-10-02 03:49:05.135 [CRT] SRVR: Health check: chain backend failed after 3 calls
2024-10-02 03:49:05.135 [INF] SRVR: Sending request for shutdown
2024-10-02 03:49:05.135 [DBG] HLCK: Health check: chain backend max attempts failed, monitor exiting
2024-10-02 03:49:05.135 [INF] LTND: Received shutdown request.
2024-10-02 03:49:05.136 [INF] LTND: Shutting down...

As you can see, my testnet bitcoind node failed... it failed in an unusual way -- the testnet bitcoind, running a docker container, suddenly failed with this error:

bitcoind Fatal LevelDB error: Corruption: block checksum mismatch:

corrupt data of some kind... right? So LND quit when it decided that the back end was down.....

.... So, I then deleted the testnet3 installation, restarted bitcoind, and allowed bitcoind to sync for a few hours.

Once bitcoind was ready, I restarted litd... starting from line 16161 in the logs....

2024-10-02 21:05:49.769 [INF] LITD: LiT version: 0.13.4-alpha commit=v0.13.4-alpha
2024-10-02 21:05:49.769 [DBG] STAT: Setting the accounts sub-server as enabled
2024-10-02 21:05:49.770 [DBG] STAT: Setting the faraday sub-server as enabled
2024-10-02 21:05:49.770 [DBG] STAT: Setting the loop sub-server as enabled
2024-10-02 21:05:49.770 [DBG] STAT: Setting the pool sub-server as enabled
2024-10-02 21:05:49.770 [DBG] STAT: Setting the taproot-assets sub-server as enabled

..LND does quite a lot of stuff, until, at line 23423, LND shuts down:

2024-10-02 22:00:28.876 [DBG] LNWL: ChannelPoint(48073b7767573658333a14956e2109f74331498ea4686a4c3340914289d37173:0): Restoring 0 dangling remote updates
2024-10-02 22:00:28.877 [DBG] LNWL: ChannelPoint(48073b7767573658333a14956e2109f74331498ea4686a4c3340914289d37173:0): Restoring 0 local updates that the peer should sign
2024-10-02 22:00:28.878 [INF] LNWL: Generating new musig2 sig for session=86e6dbbc93c5bf6ca7d6d99016f1d5b835c3ae1bfd198b20ef95da988d5113cb, nonces=NoncePair(verification_nonce=027e33f736b59f5f4a64745516791720807ac29d2f4ce305b9026cc1019ae641f303cf642b6f8f6f9a17c6c7af53cf06f581da72df6a6ad165f87ac3ad856a209827, signing_nonce=0294e3af5205b376ab8612283c43b1e6bd46b09e7954260113f8e79ceaa8c2b5680226270337ec21e4885ea8dd7e6c9ca97ac04f20215a5a6bc94d7d955d2b6ae8ef)
2024-10-02 22:00:28.879 [ERR] CNCT: ChannelArbitrator(48073b7767573658333a14956e2109f74331498ea4686a4c3340914289d37173:0): unable to force close: unable to combine musig partial sigs: error combining partial signature: error combining partial signature: final signature is invalid
2024-10-02 22:00:28.879 [ERR] CNCT: ChannelArbitrator(48073b7767573658333a14956e2109f74331498ea4686a4c3340914289d37173:0): unable to advance state: unable to combine musig partial sigs: error combining partial signature: error combining partial signature: final signature is invalid
2024-10-02 22:00:28.879 [INF] CNCT: ChainArbitrator shutting down...

I think the most important error message here might be:

unable to advance state: unable to combine musig partial sigs: error combining partial signature: error combining partial signature: final signature is invalid

You can then see that I restart litd again at line 23535....

And then litd shuts down again, for the same reason, at the end of the logs, at line 27035.

My (vague, probably wrong) theory is that I was running for a time with bitcoind with data in a corrupted state, and that may have "infected" LND... so even when I restored bitcoind's data, LND's database was still corrupted.....?

thanks!

[Roasbeef]

Thanks for the report! We're tracking this here: lightninglabs/taproot-assets#1138

[Roasbeef]

bitcoind Fatal LevelDB error: Corruption: block checksum mismatch:

This is very interesting. It points to some disk corruption that occurred, when then may have also affected lnd's database.

[MegalithicBTC]

Oh, looks like that this taproot assets issue lightninglabs/taproot-assets#1138 -- is my exact problem, my guess is that the LevelDB corruption is a red herring -- because I'm 99% sure I DO have force closures of taproot asset channels in this node. I will close this discussion and preserve the state of this node so , when that issue has a resolution I can see what I can do. Thanks.