# Elasticsearch 4 node cluster distributed in 2 rooms

**URL:** <https://discuss.elastic.co/t/elasticsearch-4-node-cluster-distributed-in-2-rooms/78075>\
**Category:** Elasticsearch\
**Created:** [March 10, 2017, 1:53am UTC](https://discuss.elastic.co/t/elasticsearch-4-node-cluster-distributed-in-2-rooms/78075 "2017-03-10T01:53:03Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![lecyberian](https://avatars.discourse-cdn.com/v4/letter/l/5e9695/32.png) [@lecyberian](https://discuss.elastic.co/u/lecyberian)\
**Post date:** [March 10, 2017, 1:53am UTC](https://discuss.elastic.co/t/elasticsearch-4-node-cluster-distributed-in-2-rooms/78075/1 "2017-03-10T01:53:03Z")

</div>

Hey everyone,

I have a scenario where the requirement is having total 4 node cluster of elasticsearch having 2 in Site A and 2 in Site B.

Both sites are basically two floors connected with fiber, so network latency is not an issue.

The requirement is to have fault tolerance and high availability. For example: if Site A goes down Site B remains functional and vice versa.

I wanted to be sure to get expert advice on this before i implement, there is a parameter called "discovery.zen.minimum\_master\_nodes" if i set that to 3 in this case i suppose it means it needs 3 master servers to be available to form a quorum just to make sure its not a split brain in case of site goes down, However if Site A goes down then this requirement will never meet and quorum will not be form i would assume that Cluster would not be functional. Any suggestions here please? 🙂

Also what primary and replica shard settings do you suggest ? provided that there are 5 indexes of size each "200-250GB" and growing

Thanks!

---

<div class="post-metadata">

**Author:** ![warkolm](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/warkolm/32/39224_2.png) [@warkolm](https://discuss.elastic.co/u/warkolm)\
**Post date:** [March 11, 2017, 7:15pm UTC](https://discuss.elastic.co/t/elasticsearch-4-node-cluster-distributed-in-2-rooms/78075/2 "2017-03-11T19:15:02Z")

</div>

This sort of situation is where you need to understand the tradeoffs you need to make.

1. Set it to 3 and then be prepared to restart the two nodes in the room that didn't go down to update the min master setting.
2. Have a tie-breaker node somewhere else.
3. Have the ability to spin up a 3rd node in either room to fulfil the 3 min.

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [April 8, 2017, 7:15pm UTC](https://discuss.elastic.co/t/elasticsearch-4-node-cluster-distributed-in-2-rooms/78075/3 "2017-04-08T19:15:13Z")

</div>

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.