Lost funds (FOUND)

[gus4rs]

I am doing some recover tests in the testnet and ended up with missing funds. Here's the sequence of events:

1. Had a channel, did some payments, but received it all back, approximately 0.1 tBTC
2. Saved the seed and channel.backup, wiped all files
3. Did an onchain recover, followed by an off-chain recover. I had 0.09212173 tBTC onchain funds.
4. The force close wasn't being done despite testnet purging only fess < 2 sat/vB
5. Forced it with

chantools triggerforceclose -t --peer 03345d4d00a5a65535513b739e78566c24d1f34a3f853814daf4fbb37f413fbf62@34.224.22.127:9735 --channel_point e45817116542f446d41b800f6661219feeed2e1d4e39d8115306fd39510f168a:0

6. The force close transaction 488ceb90119e243abc1983e866a0b13bef69d71a0b57a200093babd6f946972b was broadcasted

7. It was using a very low fee that'd take days (effective fee was less than 20 s/vB)

8. I used then

chantools pullanchor --sponsorinput 53e3e2834cff4aae1882b94aa7b902495681fcace56c62f3c84c1f8a62d81bb9:0 --anchoraddr tb1qjgz24jjq74srkksh5cl9zfmlq03h7vt0ynka2fgllnq2nkxn5wrsdmr6h0  --changeaddr fromseed --feerate 100 -t 

The sponsorinput was the only UTXO I had of about 0.9 tBTC

9. I signed the PSBT and broadcast it, the tx got confirmed: e6fa0732c9552539468ae59ce2997218c7e8a70b5b5bde5b1dd53e98eaaf093c

10. Finally did

 chantools sweepremoteclosed -t --recoverywindow 300 --feerate 20 --sweepaddr fromseed --publish

And it didn't do anything:

found 0 sweep targets with total value of 0 satoshis which is below the dust limit of 600

In my wallet, I have only about 0.096 tBTC, which is a little bit more than I had initially, and no pendingchannels

Money gone?

[gus4rs]

The weird thing is on step 8, the pullanchor spent the whole input I provided, to an address that my wallet doesn't recognize tb1qpxt4j9k46yqm5jf6xx4py9wd2e89067l00y2k9, that could explain why there are missing funds: I got the money from the channel (0.096 tBTC) but somehow pullanchor got rid of the sponsor input :/

[guggero]

Well then I need more information. When you say "it's not there, the xpub is", then how can I know it's the correct xpub? Where did you get it from? What's its derivation path? How many levels?

[gus4rs]

what info do you need? I got that xpub from lncli itself (lncli wallet accounts list --name default), and with that I can see the funding transactions and channel openings, so it's the correct xpub.
And no funds entered coming out of pullanchor. Is there a practical way to lookup for funds in non standard paths? I wasn't able to scan it from Sparrow, I'm afraid it always hardcodes to the sample path ...

[gus4rs]

I think I've found it. It is indeed the address on m/84'/1'/0'/0/0

[guggero]

#136 (reply in thread)

[gus4rs]

Guess there is a bug fixing to do on testnet! :)