We have cluster where due to provisioning bug we ended up creating large number of indexes with 100 shards each.
The cluster is 48 data nodes, with shard count now at 120,000. At this state we are not able to delete indexes that are not required as cluster is just timing out and throwing timeout exceptions.
We have stopped indexing to reduce the load, but still the cluster is unresponsive to any cluster level events. Queries seems to working fine though.
Is there a way to get rid of unwanted indexes now that cluster is in this state ?
Thanks for help in advance.