So we've had some major disk corruption on our hosted engine (hardware to blame), and we have taken backups. However, the hosted-engine VM will no longer boot at all, database is thoroughly corrupted, and we need to rebuild the thing. Just a sanity check on the best route:

Preamble:

VMs are still running fine - only hosted engine affected

VMs are distributed across our entire 3 node cluster

All 3 nodes are registered as hosted engine candidates

1. Do another hosted-engine --deploy on one of the existing hosts, and then restore the backup into that

2. Build a new host, deploy the hosted-engine, then restore a backup on a fresh node