We have been losing anywhere from one to three nodes at a time with Grandpa Voter Errors at random block numbers ranging from 1600 to over 183K. Thus far we have not found any configuration in the JSON or in the node subcommands, or in the hardware that solves it. This is my question: is it simply a fact of blockchain life that Grandpa voter errors will sometimes take your node down? I am asking this because right now we are focused on solving the issue (see this question to track us working the issue) but perhaps we should be focused on mitigation instead, i.e. how to get the node back up and running quickly when it strikes.
Add a comment
|
1 Answer
After some great insight from the Web3 Foundation folks on the Polkadot Discord, we have found that Grandpa Voter errors are NOT at all normal or to be expected and we have found some potential solutions! Closing this question out now to focus on the solutions. Details on solutions will be posted on How To Prevent "Grandpa Voter Error" And Randomly Deleted Keystore