# ES topology

**URL:** <https://discuss.elastic.co/t/es-topology/60469>\
**Category:** Elasticsearch\
**Created:** [September 14, 2016, 8:44am UTC](https://discuss.elastic.co/t/es-topology/60469 "2016-09-14T08:44:41Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![muhamadli302](https://avatars.discourse-cdn.com/v4/letter/m/b4bc9f/32.png) [@muhamadli302](https://discuss.elastic.co/u/muhamadli302)\
**Post date:** [September 14, 2016, 8:44am UTC](https://discuss.elastic.co/t/es-topology/60469/1 "2016-09-14T08:44:41Z")

</div>

Hi,  
I have ES 1.7, with two nodes in the cluster:

1. Master, with this config:  
`node.master: true node.data: true`  
it sits on a machine with 8 GB RAM and one TB space (I know there's not enough RAM, we're waiting for it to be purchased)
2. second node, with this config:  
`node.master: false node.data: true`  
it sits on a machine with 32 GB RAM and one TB space.

I need an advice, should I change the second node to be my master or not?  
And if I do make the change, what will happen to the data the current master node?

Thanks for your help!

---

<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:** [September 14, 2016, 8:52am UTC](https://discuss.elastic.co/t/es-topology/60469/2 "2016-09-14T08:52:42Z")

</div>

I'd make all 3 master eligible, see [https://www.elastic.co/guide/en/elasticsearch/guide/2.x/important-configuration-changes.html#\_minimum\_master\_nodes](https://www.elastic.co/guide/en/elasticsearch/guide/2.x/important-configuration-changes.html#_minimum_master_nodes)

---

<div class="post-metadata">

**Author:** ![kstaken](https://avatars.discourse-cdn.com/v4/letter/k/e68b1a/32.png) [@kstaken](https://discuss.elastic.co/u/kstaken)\
**Post date:** [September 14, 2016, 6:56pm UTC](https://discuss.elastic.co/t/es-topology/60469/3 "2016-09-14T18:56:21Z")

</div>

With only 2 nodes you have 2 scenarios.

1. Shards that may or may not exist on either node. Here it's probably better to just have a single master.

2. There's a copy of every shard on both nodes. Here it may make sense to run two masters for HA however you have to be aware of the possibility of a split brain if for some reason the two nodes get partitioned at the network level.

I'd recommend option 3, get a third node so you can set minimum\_master\_nodes to 2. Failing that if it were me I'd probably stick with a single master as cleaning up a split brain scenario wouldn't be fun.

---

<div class="post-metadata">

**Author:** ![muhamadli302](https://avatars.discourse-cdn.com/v4/letter/m/b4bc9f/32.png) [@muhamadli302](https://discuss.elastic.co/u/muhamadli302)\
**Post date:** [September 15, 2016, 8:06am UTC](https://discuss.elastic.co/t/es-topology/60469/4 "2016-09-15T08:06:43Z")

</div>

Hi, thanks for the reply.

We will get a third node very soon, and than I think I'll define it as:  
`node.master: false node.data: false`

And then I'll set minimum\_master\_nodes to 2. What do think?

---

<div class="post-metadata">

**Author:** ![kstaken](https://avatars.discourse-cdn.com/v4/letter/k/e68b1a/32.png) [@kstaken](https://discuss.elastic.co/u/kstaken)\
**Post date:** [September 15, 2016, 3:15pm UTC](https://discuss.elastic.co/t/es-topology/60469/5 "2016-09-15T15:15:12Z")

</div>

With 3 nodes you should run all of them as masters.

---

<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 5, 2017, 10:19pm UTC](https://discuss.elastic.co/t/es-topology/60469/6 "2017-07-05T22:19:58Z")

</div>