Open Giszmo opened 3 years ago
So as I was able to send out some funds after the above issue, I sent some more funds there to see if I can get those back instantly. Now the issue is reproducible 100%.
Please correct me if my assumptions are wrong:
Ran out of routes to try after 28 attempts: see 'paystatus'
means it could not find that route that worked before in the opposite direction? Does that mean that spark is missing topology information or is it more likely that nodes on the way favor one direction over the other? How can I see the used routes? I can't find it neither in Spark nor in Phoenix.I see this is probably a btcpayserver issue after all. Or https://github.com/btcpayserver/btcpayserver-docker/pull/448/files to be precise. They are behind by two years.
This issue was originally reported to BTCPayServer
Describe the bug
After trying to send funds using the spark-wallet, the web interface stops working with an "Error: Gateway Timeout" and the process spark-wallet runs at 100% CPU until I kill it.
Killing spark-wallet with SITTERM fixes the spark UI.
To Reproduce the bug
Unfortunately this issue is not 100% reproducible. It happened, I rebooted the machine as restarting the btcpayserver from the webinterface only resulted in the whole thing responding 502 and I ran into the state again. Killing only spark-wallet fixed it then and after that, sending funds worked or not but the server didn't hang again.
Your BTCPay Environment (please complete the following information):
Logs (if applicable)
Setup Parameters