Owning subdaemon lightning_onchaind died #1290
Comments
|
This is a Should Never Happen. I'll add some more debugging for this case so we can get an exact trace of what happens. |
|
Best logs will be if you do a You can either pull branch pr/1293 or wait until travis passes in 25 minutes and I'll merge into master. |
|
peer debug log |
|
2018-03-28T01:49:11.427Z lightningd(12461): lightning_onchaind-02ed613f005d8dae7e7211d628c63c963c4df35d9b56afabac39fefe9e90a89c75 chan #4: grind_feerate failed from 1015 - 4028 for tx 02000000017698fdb6ad5b12f964b1e721942d0f5ae8cff290da3ae9cb9c1e92ff51b9184c02000000000000000001a30a030000000000220020151afb44dd24be5f287b1399dd0cf63c62c0fde90e587647b1c07849d6dc65a80dab1300, signature 3045022100c484336113327e47ed0d4fdaba0eca49f117e98a8ae1c584929bda86a4e2eb7d0220452c93179d6c2e35234674cde56c60f3c044c88b03f5000696357826a60082cc, multiplier 663 |
|
updated listpeers debug log with grindfee output. |
We previously tried to use the commitment tx to create an initial feerate range, then refine it as we look at each HTLC tx. This was wrong, because the commitment tx can underpay fees (if it can't afford it), and our estimate of the maximum possible feerate would be too low. Now, we only have two fees we need to figure out: HTLC timeout txs and HTLC success txs, so simply grind them on first use. Fixes: ElementsProject#1290 Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We previously tried to use the commitment tx to create an initial feerate range, then refine it as we look at each HTLC tx. This was wrong, because the commitment tx can underpay fees (if it can't afford it), and our estimate of the maximum possible feerate would be too low. Now, we only have two fees we need to figure out: HTLC timeout txs and HTLC success txs, so simply grind them on first use. Fixes: ElementsProject#1290 Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
|
errors now changed slightly. 2018-04-04T08:19:13.623Z lightningd(25977): lightning_onchaind-02ed613f005d8dae7e7211d628c63c963c4df35d9b56afabac39fefe9e90a89c75 chan #4: STATUS_FAIL_INTERNAL_ERROR: htlc_timeout_fee 1672 failed sigcheck for tx 02000000017698fdb6ad5b12f964b1e721942d0f5ae8cff290da3ae9cb9c1e92ff51b9184c02000000000000000001bb06030000000000220020151afb44dd24be5f287b1399dd0cf63c62c0fde90e587647b1c07849d6dc65a80dab1300, signature 3045022100c484336113327e47ed0d4fdaba0eca49f117e98a8ae1c584929bda86a4e2eb7d0220452c93179d6c2e35234674cde56c60f3c044c88b03f5000696357826a60082cc, wscript 76a91436b3f906c15b44d86f60db0f902c4d0cdc9751c68763ac672102a7afa2906363ee562ce1b1bc73f2094a3621af05370345bb3cb73de1b717dd277c820120876475527c2103a9de3e5d78f3b253a42421c36c29ce9e8c78ffdabd061c2c8d1a5fb3e083d4a252ae67a914ab198fefb1c023a0df97a00af41a1407a9ce3a7f88ac6868 |
|
update on merge progress? |
|
The original issue was supposedly closed in a merged pull request #1293. Maybe open a new issue? |
|
#1293 only merged the changes to add more logging to see what was happening, |
We previously tried to use the commitment tx to create an initial feerate range, then refine it as we look at each HTLC tx. This was wrong, because the commitment tx can underpay fees (if it can't afford it), and our estimate of the maximum possible feerate would be too low. Now, we only have two fees we need to figure out: HTLC timeout txs and HTLC success txs, so simply grind them on first use. Fixes: ElementsProject#1290 Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Issue and Steps to Reproduce
Hard to say how to reproduce unfortunatly but this channel is stuck and cannot close due to daemon died.
getinfooutput{ "id" : "0392419c02fbca83ae8d541cd2f48b662b2238ed8e1e92a93547c23ba32f0d64c6", "port" : 9735, "address" :
[
{ "type" : "ipv4", "address" : "203.86.204.88", "port" : 9735 } ], "version" : "v0.5.2-2016-11-21-2422-g286cfef", "blockheight" : 1289331, "network" : "testnet" }
getlog
http://paste.ubuntu.com/p/bsN2tyN35X/
The text was updated successfully, but these errors were encountered: