# No replicas on .scripts index? **URL:** \ **Category:** Elasticsearch\ **Created:** [June 12, 2015, 2:10pm UTC](https://discuss.elastic.co/t/no-replicas-on-scripts-index/2579 "2015-06-12T14:10:44Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![jpotisch](https://avatars.discourse-cdn.com/v4/letter/j/c77e96/32.png) [@jpotisch](https://discuss.elastic.co/u/jpotisch)\ **Post date:** [June 12, 2015, 2:10pm UTC](https://discuss.elastic.co/t/no-replicas-on-scripts-index/2579/1 "2015-06-12T14:10:44Z")
We run rspec tests after provisioning to confirm Elasticsearch is configured properly. We do not explicitly set the value for `index.number_of_replicas` in the config, so it should default to 1. For all indices created by us or Marvel, that is what we observe. But I noticed that the `.scripts` index is always created with 0 replicas. Is this correct behavior? It seems like it should follow the global setting unless there is a separate setting specifically for `.scripts` that I'm not aware of. If it _is_ correct behavior, a) it should be documented, and b) why? It seems like the fact that `.scripts` reports one shard and no replicas means that there will always be one node that can take down our cluster, as all our searches use templates stored in the `.scripts` index. Currently I have to special-case our cluster validating logic to allow `"number_of_replicas": "0"` for `.scripts` and 1 for everything else. Any guidance on whether or not this is a bug in Elasticsearch, whether I should explicitly set `number_of_replicas` to 1 via Index Update API to ensure cluster availability if I lose my `.scripts` node, etc. would be greatly appreciated. Thanks, -joel ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [July 6, 2017, 12:07am UTC](https://discuss.elastic.co/t/no-replicas-on-scripts-index/2579/2 "2017-07-06T00:07:50Z")