Need to understand more about the Leaf node fail over scenario and how the storage distribution is happening . For example , Total 8TB online storage created 4 leaf node with 2TB each . If one leaf node failure then how the additional storage distribution is happening in the remaining 3 leaf node based on the replica . Configured “leaf_failover_fanout” is "Load Balanced "
Hi, This doc explains SingleStore HA in detail (High Availability · SingleStore Documentation)
But in summary, all SingleStore clusters have inbuilt high availability. Each leaf (at minimum) will have a primary and secondary partition. Primary partition is your master and secondary partition is your replica. These partitions are evenly spread across all your leafs, where a replica is never on the same leaf as the primary. This way when a leaf fails, replica is immediately promoted as a primary and node repair process in the background will spin up a new leaf and hydrated with current data. Because data partitions are evenly spread across multiple leafs, typically application performance is not heavily impacted when a leaf is down.
I also have facing the same problem.