# To multi path.data or hardware RAID or not?

**URL:** <https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143>\
**Category:** Elasticsearch\
**Created:** [January 26, 2016, 5:15pm UTC](https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143 "2016-01-26T17:15:10Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![javadevmtl](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/javadevmtl/32/45613_2.png) [@javadevmtl](https://discuss.elastic.co/u/javadevmtl)\
**Post date:** [January 26, 2016, 5:15pm UTC](https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143/1 "2016-01-26T17:15:11Z")

</div>

So I have my servers with RAID 0 over 6 drives. We lost a disk, no sweat we had replicas in the cluster.

Just thinking should I try setting up using multiple paths?

Thinking of creating 3 RAID 0 drives and then setting up ELasticsearch to use those 3 paths.

Is that a good idea? If we had done this earlier. We would have only lost that 1 RAID only those shards correct, bit not the whole 6 disks?

---

<div class="post-metadata">

**Author:** ![javadevmtl](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/javadevmtl/32/45613_2.png) [@javadevmtl](https://discuss.elastic.co/u/javadevmtl)\
**Post date:** [January 26, 2016, 10:35pm UTC](https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143/2 "2016-01-26T22:35:05Z")

</div>

Ok so I went down this route...

I created 3 separate logical disks using RAID0 from the 6 disks. And I configured ES to have 3 data paths. Seems to be working Ok so far.

---

<div class="post-metadata">

**Author:** ![jasontedor](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/jasontedor/32/66992_2.png) [@jasontedor](https://discuss.elastic.co/u/jasontedor)\
**Post date:** [January 27, 2016, 1:46am UTC](https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143/3 "2016-01-27T01:46:42Z")

</div>

> [@javadevmtl](#):
>
> Just thinking should I try setting up using multiple paths?

It depends. 🙂

The tradeoffs are pretty straightforward. Multiple paths in `path.data` stripes your shards across the paths (at the shard level), and using RAID 0 stripes your shards across the disks (at the block level). In the first case if you lose a disk, you lose the shards on that disk and nothing else. In the second case, if you lose a disk, you lose all of your shards on that machine. Therefore the first gives you increased safety, but the trade off is lower performance.

It depends on _your_ workload whether or not you need the increased performance, and whether or not you have the risk appetite to accept the reduced safety that comes from choosing the increased performance.

---

<div class="post-metadata">

**Author:** ![javadevmtl](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/javadevmtl/32/45613_2.png) [@javadevmtl](https://discuss.elastic.co/u/javadevmtl)\
**Post date:** [January 27, 2016, 10:22pm UTC](https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143/4 "2016-01-27T22:22:11Z")

</div>

Yeah that's fine. Testing with a hybrid approach.

---

<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, 11:20pm UTC](https://discuss.elastic.co/t/to-multi-path-data-or-hardware-raid-or-not/40143/5 "2017-07-05T23:20:56Z")

</div>