Describe the bug
After a block is mined, bitcoin-core does not immediately save it to disk.
But will broadcast it to clients. (in this case lightning nodes)

If bitcoin-node does not get cleanly stopped/restarted, that unpersisted block will be lost.
And when some client with previous block header tries to query it with blockhash, it will fail with "BlockNotFound"

To Reproduce
Steps to reproduce the behavior:
Mine a block then kill container or stop network.

Solution

• Polar must stop bitcoin d via rpc stop in shutdown script, so that it persists and flushes all data to disk before exiting.
• It also needs to be handled in clients consuming bitcoin blocks.