@servint All hardware can fail. The question is not how to prevent failures, but how to ensure failures won't affect service. Ironic now.
