Fix .script index template.

[GaelTadh]

This commit makes the default number of shards for the .scripts index to 1, it also
forces the auto_expand replicas to 1-all. This change means that script index GET requests to load
scripts from the index should always use the local copy of the scripts index, preventing any network traffic or calls on script GET.

[s1monw]

@GaelTadh this looks good to me I think @martijnvg should look over this one more time as well.

[martijnvg]

@GaelTadh LGTM

[sofianito]

Is this issue fixed in 1.3.3 and 1.3.4?

Thanks

[clintongormley]

yes

[sofianito]

In 1.3.2, the .scripts index was created by default with 5 primary and 1 replica shards.

I just upgraded to 1.3.4, and now the .scripts index is created with 1 primary and 0 replica shards.

The only way I figured out to change the .scripts shards number is to create it manually before indexing any search template:

PUT .scripts
{
  ".scripts": {
    "settings": {
      "index": {
        "number_of_replicas": "1",
        "number_of_shards": "5"
      }
    }
  }
}

Is this the recommended way? Are there any recommended values for the .scripts' replicas and shards numbers?

Thanks.

[clintongormley]

@sofianito that is the recommended way, but why do you want so many shards for scripts? one shard should be plenty

[sofianito]

Thank you very much!. I'll ask our Devops team to stick with default value.

[sofianito]

@clintongormley: One more question. What happens if we have a cluster of two nodes and the master goes down? I understand we won't have failover since the .scripts default replica number is 0. Which means we would need to change the default replica number to 1 at least, isn't it?

Thanks

[clintongormley]

@sofianito no - if you read the description at the top, you'll see that auto_expand_replicas is set to 1-all, so there should be a copy of this index on every node.