# Syncing two clusters

**URL:** <https://forum.opensearch.org/t/syncing-two-clusters/15242>\
**Category:** Request For Comments\
**Tags:** discuss\
**Created:** [July 25, 2023, 9:01am UTC](https://forum.opensearch.org/t/syncing-two-clusters/15242 "2023-07-25T09:01:11Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![szurokl](https://avatars.discourse-cdn.com/v4/letter/s/a5b964/32.png) [@szurokl](https://forum.opensearch.org/u/szurokl)\
**Post date:** [July 25, 2023, 9:01am UTC](https://forum.opensearch.org/t/syncing-two-clusters/15242/1 "2023-07-25T09:01:11Z")

</div>

Hello OpenSearch community,

I would like to ask what is the current recommended way to synchronize/backup indices between two OpenSearch clusters? (The more frequent the synchronization can be, the better + I would like to be able to automatically delete old data)

I know CCR exists, but I run into problems with it (auto deletion is not a part of the plugin yet, ISM policy unable to delete replicated index, due to it being readonly, CCR jobs get stuck in the background trying to replicate non-existent indices).

I considered using [elasticdump](https://github.com/elasticsearch-dump/elasticsearch-dump), but it seems to be kind of slow, or configuring logstash with separate pipelines to write to multiple clusters.

If anyone has experience with similar use case, I would welcome some recommendations.

---

<div class="post-metadata">

**Author:** ![Gsmitt](https://sea1.discourse-cdn.com/flex019/user_avatar/forum.opensearch.org/gsmitt/32/1718_2.png) [@Gsmitt](https://forum.opensearch.org/u/Gsmitt)\
**Post date:** [July 27, 2023, 2:37am UTC](https://forum.opensearch.org/t/syncing-two-clusters/15242/2 "2023-07-27T02:37:53Z")

</div>

Hey @szurokl

> [@szurokl](#):
>
> synchronize/backup indices between two OpenSearch clusters?

Since you stated two clusters, you could configure **Data Sources** shown [here](https://opensearch.org/docs/latest/dashboards/discover/multi-data-sources/).

Basically connect to another opensearch/elasticsearch node. once completed create a index pattern, then pretty much create anything from that data on one node.

I havent tried to sync indices yet, but you could configure snapshot policies, Make sure you created a index pattern then create you snapshot policy from that. Not sure if it will work, I have not tried it yet.

 ![image](https://us1.discourse-cdn.com/flex019/uploads/mauve_hedgehog/original/2X/6/6ec5c835d73afe69b47a62dd74d9215a2e2573ff.png)