Redis cluster inservice upgrade is not working with zero downtime

[TANISH-18]

While doing the Redis cluster in-service upgrade from older chart to newer chart. Redis operator is performing the Rolling update shard by shard which is fine.
but in one shard it is upgrading the master replicas at once due to which there is disconnection with the application.

can we improve the Rolling update logic such that first it upgrade the replicas of one shard then promote that replica to master and then we upgrade master and new master will join back as replica.?

[TANISH-18]

(

operator-for-redis-cluster/pkg/controller/actions.go

Line 122 in 24f4e42

func (c *Controller) manageRollingUpdate(ctx context.Context, admin redis.AdminInterface, cluster *rapi.RedisCluster, rCluster *redis.Cluster, oldNodes, newNodes redis.Nodes) error {

)