We've been running through a replay of the EOS chain (with the state history plugin and mongodb plugins enabled).
I had seen elsewhere that a SIGTERM (i.e.
kill <pid>) would safely stop nodeos, but stopping this way has still resulted in a dirty database / need for a full replay when trying to restart nodeos.
We're using nodeos v1.5.0.
What is the correct way to stop the nodeos process so that it can be resumed cleanly without the risk of a database dirty flag?