We tried to figure out a way to restore a degraded A-A Cluster without shutting the cluster down (or at least with minimum downtime). We want to use the A-A Cluster as a kind of RAID 10 between two servers to avoid double redundancy (and double costs for the SSDs). For this and for better performance we use RAID 0 configurations inside the single Server which mirrors with the second server in a A-A Cluster. We are looking for the same kind of "plug and play" comfort and the easy handling that you have with a "normal" hardware disk-RAID 10 or 5, which means that you can drag the broken disk out and replace it with a new one without shutting the RAID down.

If this happens with the RAID 0 inside one of the A-A partners (which will be more likely with a RAID 0) or if one of the servers gets totally broken and needed to be replaced we have to restore the complete configuration to the exact same status it had before this accident and re sync the two servers again. Is there a best practice how to do this ?