Closed silenzara closed 1 year ago
Yea, sadly its kinda intended - we have to limit our total max dust exposure at any one time to ensure we're not vulnerable to dust-inflation attacks by a mining counterparty (I believe all lightning implementations do this in one way or another). That said, with current feerate spikes dust issues are too easy to hit today. #2261 talks about how we need to bump the default dust exposure threshold config knob, which might have saved you here, and further this is not a reachable error in the post-anchors world, which we also intend to ship in the next release.
Sadly in the pre-anchors lightning protocol there's not a lot we can do to address this, aside from making it more rare and migrating to 0-fee-htlc-anchors.
Thanks for the explanation, makes sense.
I thought it was related to the other dust issues but since I'm not familiar enough with LDK I wasn't sure this was the same thing. At least this specific issue will come up if someone searches for the specific error I got (Peer sent update_fee with a feerate...).
Cheers.
I just had a channel force close on me which started after my LND 0.16.2 node tried to update the channel commit fee which resulted in an error:
After this the channel link could not be reestablished, logs with relevant lines here: https://pastebin.com/raw/AiurbjkD
I'm not sure if this behavior is intended and since I could not find any issues related to "over-expose us to dust-in-flight on our own transactions" I'd thought to report it here.
Closing tx: https://mempool.space/tx/1521dfe44106225d2aa7d3cd45df6128eab55daa05444ab386bdf4c2d3a708c7
A 11,638 sat HTLC was involved as you can see in the closing tx.