Datanode computer failure (3 replies)

I have spent a lot of time searching and can't find this. If I have missed it, please forgive me.

Assume I have a cluster with at least 2 datanodes (my actual one has four but I don't think it matters). Now assume that one of the computers that contains a datanode goes down and cannot be restarted. Truly the computer or disk drives are dead and will never come back.

However, I have people that read and write from the database 24x7 and I can't shut down the entire cluster.

Is there a way to replace the dead computer/datanode with a different machine and get redundancy back without shutting down the cluster.

Thanks in advance for any help,
--Bruce