We looked through the logs, which show a crash or hard kill on 2017-02-11 15:40:58. Is it possible that you killed the multichaind process without issuing the stop API command, and then at some point restarted it? If so we're still aware there are some of these "disgraceful shutdown" issues, and they're on the to-do list. The way to recover from this kind of thing is to restart multichaind with the -reindex flag (or in this case, even -rescan should be sufficient).
If the multichaind process crashed, rather than you killing it yourself, please let us know.