Closed ManfredKarrer closed 9 years ago
OK, we got this sorted out. The problem was the server had got itself into some state where it didn't notice one of the pledges was revoked, and kept serving it, so client tried to use it and then the contract was seen as an orphan because one of its utxos was missing. The pledge in question was created by a Lighthouse that crashed half way through .... it shouldn't have been an issue but I guess the combination of two problems resulted in issues.
The server had been running for quite a long time. I wonder if it was stuck somehow. Unfortunately debugging infrastructure on the testnet sux .... the only block explorer that works on it doesn't show you the connections between transactions properly or much other technical info, so it's a bit hard to figure out what happened.
I'll investigate the server to see why it got stuck this way.
A potential candidate for a Lighthouse project (better tesnet environment)?
We also suffer from the testnet environment for Bitsquare testing, specially because confirmation time is often 20 min. (don't know the reason for that). To have an alternative testnet with a short confirm time (or at least more mainnet-like) just for app testing purpose (I know it must survice chaotic testnet as well, but often you dont need that stress test situation in development) would be really nice to have. Of course I know RegTest and use it but at some stages you need WAN side network...
I haven't seen this since. If we spot it again, let's reopen.
After stresstesting a project (many pledge/revoke, overpay - revoke) i wanted to claim the 100% funded project but it hangs in the claim pledges screen. when i had my local bitcoin core running it the progress indicator was stopping exactly at 50% without the bitcoin core it goes one step further but does not finish as well. I will send you the logs by mail.