-
Notifications
You must be signed in to change notification settings - Fork 912
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
lightningd stuck while funding a channel with a custom backend and eat memory around 20 GB of RAM #6183
Comments
Ok looks like that with wrong (or too big estimation fee |
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
I cannot reproduce this. See #6204 |
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: ElementsProject#6183 Signed-off-by: Rusty Russell <[email protected]>
See: #6183 Signed-off-by: Rusty Russell <[email protected]>
We just hit this in production and our node is unusable. @vincenzopalazzo is the solution just to upgrade to the latest version? |
Maybe I never trigger the bug again, but if this happens means that bcli plugin is going crazy. I see also you PR, did you solve the problem of upgrading to the new bitcoin version? |
We're still getting infrastructure updated across the board and will find out soon. Though we aren't doing the same thing as you were by running a custom backend that was providing fee rates, so I highly doubt this will fix our issue and there's some other critical bug with the |
I see keep me updated so we can debug it |
@vincenzopalazzo It's still happening with 23.08:
Further, CLI commands for everything, including We are going to do a chain resync now, but any guidance on how to proceed? |
For ease, here's some of the code for each of the parts in the stack trace: Lines 553 to 605 in 98ad467
Lines 221 to 274 in 98ad467
Lines 513 to 533 in 98ad467
And then it goes into This has not been a problem all year and just surfaced yesterday. There's maybe 1k+ channels it has had. When the lock up happens, it behaves exactly like described in the OP. It eats up RAM slowly until it runs out. |
I need to look more at the stacktrace, I will do in a couple of hours, but this is a crash caused by a OOM? or just a a crash? I am thinking that this can be a bug introduced in this release? |
@vincenzopalazzo There is no real 'crash'. CLN just completely freezes up and stops responding. Then its memory consumption grows until there is some OOM kill situation. This was happening on v23.05.1 and v23.08 |
Thanks to confirm so this is the same bug, putting in the point release queue lets see if we can debug it thanks |
I've reproduced it and got a fix at #6657 The temp solution I think is to fund the node with more higher value UTXOs, but we are still doing mantenance on our node so we have not verified it yet. |
While testing a new backend plugin for core lightning, and testing it I get stuck with the
fundchannel
command.Originally I thought that this was a problem from my plugin side coffee-tools/folgore#21 but then I make a stack report from core lightning with
kill -SEGV <PID>
and then I get the following reportand core lightning is stuck with the following command
Putting this there because the stacktrace in
wallet_find_utxo
sounds worrying to me!The text was updated successfully, but these errors were encountered: