# Write consistency quorum

**URL:** <https://discuss.elastic.co/t/write-consistency-quorum/25018>\
**Category:** Elasticsearch\
**Created:** [July 7, 2015, 7:40am UTC](https://discuss.elastic.co/t/write-consistency-quorum/25018 "2015-07-07T07:40:34Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![usualstuff](https://avatars.discourse-cdn.com/v4/letter/u/b3f665/32.png) [@usualstuff](https://discuss.elastic.co/u/usualstuff)\
**Post date:** [July 7, 2015, 7:40am UTC](https://discuss.elastic.co/t/write-consistency-quorum/25018/1 "2015-07-07T07:40:34Z")

</div>

Hello.

We currently have setup with multiple virtual machine hosts and two SANs for failover.  
I was thinking about best way to deploy cluster in this configuration.

So I thought about deploying some nodes with master role for split-brain protection and make other nodes serve data only. Question I've been thinking of: is it possible to override default quorum value for write consistency so cluster may write data if one of SANs goes down?  
I know about "discovery.zen.minimum\_master\_nodes" for master quorum and been searching for something like this for write consistency quorum. Another way to do this is just do not use quorum at all but I don't think this is optimal solution.

Thanks,  
Alex

---

<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:** [July 8, 2015, 4:20am UTC](https://discuss.elastic.co/t/write-consistency-quorum/25018/2 "2015-07-08T04:20:54Z")

</div>

There is write quorum, but how can ES know if a SAN has gone down?

---

<div class="post-metadata">

**Author:** ![usualstuff](https://avatars.discourse-cdn.com/v4/letter/u/b3f665/32.png) [@usualstuff](https://discuss.elastic.co/u/usualstuff)\
**Post date:** [July 8, 2015, 6:11pm UTC](https://discuss.elastic.co/t/write-consistency-quorum/25018/3 "2015-07-08T18:11:06Z")

</div>

Hello, Mark.

So the best option would be just wait until missing replicas will be allocated on live part of the cluster?

---

<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:** [July 8, 2015, 10:07pm UTC](https://discuss.elastic.co/t/write-consistency-quorum/25018/4 "2015-07-08T22:07:10Z")

</div>

And use allocation awareness as well.

---

<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 6, 2017, 12:02am UTC](https://discuss.elastic.co/t/write-consistency-quorum/25018/5 "2017-07-06T00:02:45Z")

</div>