Tapicero should log verbosely when process is dying

there's "tapicero doesn't start after first deploy to single node (Bug #6004 (closed))" where tapicero dies for unknown reason, and there's no trace in the logs about it. this makes it hard to investigate. don't know how complex it would be to get some crash reporting into the logs.

(from redmine: created on 2014-08-25, closed on 2016-01-05, relates #6004 (closed), relates #6537 (closed))