Syncthing was in the middle scanning of one of my larger sets and it restarted (or crashed and restart.)
Now, it doesn’t start and the last lines in the log is:
INFO: Database block cache capacity 65536 KiB
FATAL: Cannot open database: leveldb/table: corruption on table-footer (pos=1970978): bad magic number [file=773017.ldb] - Is another copy of Syncthing already running?
Are you sure you don’t have hardware failures (check dmesg)?
I’ve noticed quite some issues lately, maybe it’s time we should push the bolt branch a bit further?
I do think we need a way to migrate the database. This is because if the clients are not in sync, all differences will result in a conflicting file. This is especially problematic when devices contain ‘old’ data.