Closed georgiy-belyanin closed 4 months ago
Instead of the extra failover
configuration section I've implemented an approach with the extra optional shading replica configuration property.
return {
sharding = {
['cbf06940-0790-498b-948d-042b62cf3d29'] = {
replicas = {
['8a274925-a26d-47fc-9e1b-af88ce939412'] = {
name = 'storage_1_a',
master = true,
failover_priority = 1
},
['3de2e3e1-9ebe-4d0d-abb1-26d301b84633'] = {
name = 'storage_1_b',
failover_priority = 2
},
['001688c3-66f8-4a31-8e19-036c17d489c2'] = {
name = 'storage_1_c',
failover_priority = 3
}
},
}
}
}
In my opinion it's better not to manage with the duplicates of the replica sets and replicas UUIDs and to specify the failover priorities near the definition.
Also I haven't found the corresponding issue yet.
Add the
failover_priority
config option in the]sharding.<replicaset_id>.replicas.<replica_id>
section.Using the option you may specify the order failover will connect to the replicas. The lower the value, the higher the priority of the replica.