Closing transaction stuck in mempool with low fee about lightning

[vincenzopalazzo]

Sorry to be late here, what is the sequence of commands that you follow to do CPFP?

[Lagrang3]

It's a force close, by the message in the listpeers command. The outputs are Segwit scripts (P2WSH) that's why the address doesn't show in your wallet.

[guybrushino]

Sorry to be late here, what is the sequence of commands that you follow to do CPFP?

As in this case here, I simply spent my output at a higher fee with lightning-cli withdraw
https://mempool.space/tx/452fd94859200ecc5da3ab5508cd70daf665e5ada603824f720f9fa762cfc354

But in the case of a force close, whose peer has disappeared (i think forever) and whose fees have been too low for more than 4 months, what can be done?

[vincenzopalazzo]

Ah good point @Lagrang3

I ingore the json output, I assume that this is a pre-anchor channel, and I need to think a bit about how it is possible to replace this TX, when I will be back to my desk at the end of the week, but maybe @Lagrang3 has some idea

[Lagrang3]

This is precisely why we needed the anchors. If you broadcasted the force close tx, your lightning node cannot spend your output before the relative timelock, with respect to the blockheight when the transaction was confirmed, but since it has not been confirmed, the "time" is frozen. The counterparty (your peer) doesn't have the timelock in his output, he is the one who can CPFP.

[guybrushino]

Ok, I think I understand that nothing can be done, except hoping that the mempool returns to 2 sat/vbyte 🙂
(or paying a miner to mine my commitment tx)

[Lagrang3]

Yes, either:

1. you pay out-of-band, or
2. transations fees return to 2 sat/vbyte, or
3. your peer recovers his node and bumps the tx to confirm this 0.01 BTC he has available there, or

what if your peer completely lost his node's database and hdwallet?
and your node has not deleted old commitment transactions.
Maybe, there's an older commitment tx with a higher feerate.
You could risk broadcasting that tx, which will effectively RBF the current tx.

In the worst case your peer will wake up and punish you taking all the funds which where anyways mostly on his side.
And if he doesn't it probably means that he lost his data and he wasn't unable anyways to recover the funds.

[guybrushino]

Interesting!
Is it possible that I find an old commitment transactions in the CoreLn database?
Or every time it is renegotiated, the previous transaction is deleted?

[guybrushino]

Ok, I understand from this answer by @cdecker (https://bitcoin.stackexchange.com/a/96886) from 2020 that Core Lightning purposely doesn't keep old commitment transactions.

So I can't do much about it.

"Core Lightning doesn't store old versions of the commitment transaction, in order to avoid accidentally cheating."