I've played a bit with snap/restore, and noticed that when restoring with replicas, the cluster chooses a node to be a master, restore to it in a few seconds, and then starts to replicate to all other nodes. Is there a way to tell elastic to restore the to all nodes and then to choose a master ? The reason I am asking is because all nodes see the same mount, and can fetch the data, while when replicating it from the master node, it slows the master node for a long time.
Index size 6G. number of data nodes -> 26.
Thanks.