# AWS EBS recommendations

**URL:** https://discuss.elastic.co/t/aws-ebs-recommendations/108461
**Category:** Elasticsearch
**Created:** [November 20, 2017, 9:04pm UTC](https://discuss.elastic.co/t/aws-ebs-recommendations/108461 "2017-11-20T21:04:06Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![Rajesh\_Kodali](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/rajesh_kodali/32/53882_2.png) [@Rajesh\_Kodali](https://discuss.elastic.co/u/Rajesh_Kodali)
#### Post date: [November 20, 2017, 9:04pm UTC](https://discuss.elastic.co/t/aws-ebs-recommendations/108461/1 "2017-11-20T21:04:06Z")

</div>

We currently use m4.x2large instances running data nodes. I saw elastic recommends ephemeral instances. As m4.2xlarge doesn't offer instance disks, should we swith to another instance type?

---

<div class="post-metadata">

### Author: ![mujtabahussain](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mujtabahussain/32/17514_2.png) [@mujtabahussain](https://discuss.elastic.co/u/mujtabahussain)
#### Post date: [November 20, 2017, 10:46pm UTC](https://discuss.elastic.co/t/aws-ebs-recommendations/108461/2 "2017-11-20T22:46:30Z")

</div>

Just my $0.02 in that we use `i3`s for our data nodes for the NVMe disks that they offer.

---

<div class="post-metadata">

### Author: ![rudolphk](https://avatars.discourse-cdn.com/v4/letter/r/b9e5f3/32.png) [@rudolphk](https://discuss.elastic.co/u/rudolphk)
#### Post date: [December 14, 2017, 4:50pm UTC](https://discuss.elastic.co/t/aws-ebs-recommendations/108461/3 "2017-12-14T16:50:14Z")

</div>

How do you deal with the ephemeral storage of the I3 instances? I though the data gets wiped out if the node goes down?

---

<div class="post-metadata">

### Author: ![mujtabahussain](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mujtabahussain/32/17514_2.png) [@mujtabahussain](https://discuss.elastic.co/u/mujtabahussain)
#### Post date: [December 14, 2017, 10:30pm UTC](https://discuss.elastic.co/t/aws-ebs-recommendations/108461/4 "2017-12-14T22:30:25Z")

</div>

Nodes do not get wiped out willy nilly 🙂 ... But even if they do, remember you have a cluster! So when a node goes down, the cluster rebalances and you have data! This is where organising your cluster to cater to such scenarios is important.

Also snapshots! Backing up the data outside the cluster frequently means you won't lose everything even if your entire cluster goes down.

---

<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: [January 11, 2018, 10:30pm UTC](https://discuss.elastic.co/t/aws-ebs-recommendations/108461/5 "2018-01-11T22:30:38Z")

</div>

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