You must take care of waiting a few seconds before copying: bitcoind takes a few seconds to stop & correctly close the database. If you send the kill signal then directly copy the directory, it is possible the database is still marked as open and the leveldb library will refuse to open it, leading to the exception you're hitting. Please add a "sleep 15" after your kill, and check your bitcoind is really down.
If that does not do it, the best I can do is to show the leveldb error message instead of the exception; It would give you an hint. I'll take a look about this tonight.
I have started the bitcoind, waited it to synchronize with network and then stopped via
bitcoin-cli stop (this is the correct way as per docs). After a few minutes I copied chainstate and run your software - unfortunately, it still dying with the same error. So it seems like there some problem with leveldb library - maybe it is too old (1.12.0) and incompatible with bitcoin core 0.16.0 database.
Ok I will be waiting and monitoring your git for the version which catches leveldb error.
Thank you for your help. I appreciate it.