1- For BlueStore, It wasn’t clear for me how it works. Does it create LVMs each contains multiple disks?
2- Is HA option available for admin nodes? Is it provided by SES or it needs to be implemented manually (corosyns, pacemaker …)?
3- how many gateway nodes are needed in the cluster?
“1) Generally the storage admin doesnÂt need to be worried about the use of LVM. LVM is used by as a convenience for Âceph-volume to manage the physical devices. The various storage devices on a host  when BlueStore is chosen  will all belong to a common logical volume group. And depending on how the Data, WOL, and DB roles are chosen, logical volumes are established.
2) Yes, HA is an option for the admin nodes. A brief explanation is given in the course, using HAProxy, but other HA implementations are possible depending on the use case and requirements of the environment. The SES documentation has an excellent, thorough example using HAProxy.
3) Zero gateways are required for a cluster. But depending on your use case, you will deploy the desired gateways. If your users donÂt need CephFS, then you would not need to deploy any MDS Âgateways. If you only will have data provided from the cluster as Objects, then you would need at least one RGW. To provide high availability access to the RGW buckets, you would need to deploy at least 2 RGWs. To provide more isolated access to buckets  e.g.: Group A needs secure access to Bucket A, and Group B needs secure access to Bucket B, and both groups require high availability to their respective buckets  then you would need 4 RGWs.”