# Even number of nodes with garbd

**URL:** <https://forums.percona.com/t/even-number-of-nodes-with-garbd/3855>\
**Category:** Percona XtraDB Cluster 5.x\
**Created:** [November 6, 2014, 3:49am UTC](https://forums.percona.com/t/even-number-of-nodes-with-garbd/3855 "2014-11-06T03:49:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![joelf](https://avatars.discourse-cdn.com/v4/letter/j/d26b3c/32.png) [@joelf](https://forums.percona.com/u/joelf)\
**Post date:** [November 6, 2014, 3:49am UTC](https://forums.percona.com/t/even-number-of-nodes-with-garbd/3855/1 "2014-11-06T03:49:40Z")

</div>

I want to run the common setup where I have 2 data centers and need to use garbd in a 3rd. I have everything setup and working in the DC’s now I am working with garbd. I want to run garbd in EC2 so I needed to map public addresses back to the internal addresses in the DC’s to the cluster nodes. When I start garb in EC2 to the public mapped addresses everything goes into a confused state. I tried garb locally to the DC’s with internal addresses and it works as expected. Questions:

- Is there a way around the external/internal ip mapping that won’t confuse garbd/cluster
- Is it possible to run more than one garb (one in each DC that would avoid having to use a 3rd DC like EC2)

Wasn’t able to find anything specific to my problem so I figured I’d ask the community. Thanks for any info.