Open EnchanterIO opened 4 years ago
Current status
After the first research over this issue I could narrow the issue to the scenario where a node being shut down triggers tendermint dbs to close down and it appears to happen that the tendermint stateDb
is closed and Tendermint node try to persist a block right after that happens. This is a racing condition and it does not happen on every shut down.
Next steps Try to validate the potential solution of stopping Tendermint ABCI reactor before DBs are closed
Seems like the 1s timeout should be enough for persisting the changes to disk. If it fails we increase it or implement a proper solution. I am bumping the version and will test it for few days on my node. If everything goes fine we merge it by the end of the month 👍
How to reproduce
Steps to resolve: