is running, leading to need to delete and recreate the entire Fulcrum database?
One of my Raspberry Pi servers crashed due to flaky power problems with a USB disk which suddenly froze all disk io. Of course no logs. Restarted and bitcoind recovered in a few minutes, it checkpoints state. Fulcrum restart ordered up a deletion of the entire database and reloading. (This system was terribly flaky since I got a new USB storage device and has crashed perhaps 10 times, so usually the database corruption doesn’t happen.)
Nothing new here, similar problems date back to Electrum X and Electrum. But it would be nice if Fulcrum were at least as fault tolerant as bitcoind. Is anyone working on improving this?
I think it was actually a design choice. Trade-off is increased performance.
Обсуждают сегодня