So today, i had 4 pending HTLCs to 2 nodes. The first 2 expired (i think i still had a connection to the peer), and the channel is now in force close state. The last 2 are still pending until expiry and the peer seems offline, so i guess my node will trigger another force close in a few hours.

I would like to understand if it might be because of something i did wrong to avoid unneccessary force closes.

What i did recently: I opened 3 new channels in the last week, with some associated rebalance transactions. This triggered a big batch of routed payments.

During this time, the only errors i could see in lnd logs were a few failed routings because my automatic 'max_htlc' adjuster had triggered a channel update (lowering max htlc value), but some remote nodes were not yet aware of it and tried toroute bigger payments. I guess in that case, this doesn't even go to the point of creating an HTLC?

The node also got shutdown (the lnd process received an external SIGABORT) and restarted within 2 minutes. I don't know if the HTLCs were pending at that point, but if so, i guess they should have resumed and cleared by then?

The node has been online and shown no error since then

Is it more likely caused by another peer than me, or do you see something there from my side that might be the cause?

Should i make sure my max htlc adjuster never goes below the value of currently pending htlcs? I doubt the channel policy should have an effect on already created ones, but maybe i am wrong?

Should i make sure there are no ending HTLCs before closing LND, even if it is only for a few minutes?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.