We are creating large index around 90 crore documents, somehow at the time of index creation cluster health becomes red and some shard is not recoverd.
Same problem occure us 2 times, first time we are indexed document more than 80 crore and the such issue occure and we try out solutions to solve it but we didn't recover it. after that we recreate index.
and again after indexing 10 crore records same issue occures.
for this purpose we are using elasticsearch-1.6.0 version
with 6 node cluster having 1 balancer node, 1 master node, 1 master and data node and 3 data nodes.