I have been asked to restore data for a 3 node ES cluster to a new read-only cluster. The new cluster is only for showing old log data and have very few request.
I have set up one server that will be my "cluster". When I run my restore command I get 5 shards and 5 unassigned shards and I think that this is redundant as one must be enough.
How can I restore my data so I use as little disk space as possible?
Your cluster must be yellow since there are unassigned shards. Simply run the following command to remove the unassigned replica shards and the cluster will turn green again:
PUT index-name/_settings
{
"number_of_replicas": 0
}
Just note, though, that removing the unassigned replicas will not save you any disk space since those replica shards do not take up any space because they are unassigned anayway.