# Single Master with two Load Balancing Nodes

**URL:** <https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297>\
**Category:** Elasticsearch\
**Created:** [June 25, 2018, 4:56pm UTC](https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297 "2018-06-25T16:56:11Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![manoj4321](https://avatars.discourse-cdn.com/v4/letter/m/ea5d25/32.png) [@manoj4321](https://discuss.elastic.co/u/manoj4321)\
**Post date:** [June 25, 2018, 4:56pm UTC](https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297/1 "2018-06-25T16:56:11Z")

</div>

hello,  
I am trying to make a cluster in which there is one master node and the other two are data nodes (for load balancing during search query). Currently i am facing the problem in which all the shards are diverted to one of the two data nodes (checked via monitoring through kibana). When I intentionally shut down that node, I am getting the error that all primary shards are lost. But when, I shut down the other node (that too intentionally) cluster is running ok. I am running all three nodes on different machines, shared with each other. So,

Instead of two data nodes should I make two coordinating nodes ?? (As far as I know every node is a coordinating node)  
I want, if my one load balancing node breaks down the other node should handle it's work. For that what should i do ??

---

<div class="post-metadata">

**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\
**Post date:** [June 25, 2018, 4:59pm UTC](https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297/2 "2018-06-25T16:59:19Z")

</div>

> [@manoj4321](#):
>
> Instead of two data nodes should I make two coordinating nodes ?? (As far as I know every node is a coordinating node)

No, I would recommend one dedicated master-eligible node and two master/data nodes. This will give you high availability, at least as long as you set `minimum_master_nodes` to 2 and configure 1 replica for all indices.

If data is not distributed across both nodes, check that the nodes in the cluster are running exactly the same version of Elasticsearch and that they are all part of the cluster.

---

<div class="post-metadata">

**Author:** ![manoj4321](https://avatars.discourse-cdn.com/v4/letter/m/ea5d25/32.png) [@manoj4321](https://discuss.elastic.co/u/manoj4321)\
**Post date:** [June 25, 2018, 5:02pm UTC](https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297/3 "2018-06-25T17:02:28Z")

</div>

thanku @Christian_Dahlqvist for the quick reply....all the nodes are using same version of elasticsearch i.e. version 6.3 and they are part of the cluster as i am able to see all three nodes up and running while monitoring through kibana but all the shards are diverted to a single data node

---

<div class="post-metadata">

**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\
**Post date:** [June 25, 2018, 5:05pm UTC](https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297/4 "2018-06-25T17:05:08Z")

</div>

Can you provide the output of `curl -X GET "localhost:9200/_cat/nodes"` ?

---

<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:** [July 23, 2018, 5:05pm UTC](https://discuss.elastic.co/t/single-master-with-two-load-balancing-nodes/137297/5 "2018-07-23T17:05:13Z")

</div>

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