# How snapshot work and maintain?

**URL:** <https://forum.opensearch.org/t/how-snapshot-work-and-maintain/20995>\
**Category:** OpenSearch\
**Created:** [August 27, 2024, 5:06pm UTC](https://forum.opensearch.org/t/how-snapshot-work-and-maintain/20995 "2024-08-27T17:06:16Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![kitkit](https://avatars.discourse-cdn.com/v4/letter/k/c89c15/32.png) [@kitkit](https://forum.opensearch.org/u/kitkit)\
**Post date:** [August 27, 2024, 5:06pm UTC](https://forum.opensearch.org/t/how-snapshot-work-and-maintain/20995/1 "2024-08-27T17:06:16Z")

</div>

**Versions** (relevant - OpenSearch/Dashboard/Server OS/Browser):  
AWS Opensearch 7.4

**Describe the issue** :

I’d like to take manual snapshots and keep only the five most recent ones to prevent excessive storage growth. I understand that snapshots are incremental, but I’m curious about how the snapshot lifecycle works.

Scenario

- Index A with document 1
- Take snapshot 1
- Index A with document 1 and 2
- Take snapshot 2
- Index A with document 1,2 and 3
- Take snapshot 3
- and so on

I want to maintain a manageable number of snapshots, keeping only the five most recent ones. If I delete snapshot 1 and then restore snapshot 3 to the cluster, will document 1 be missing from the restored snapshot 3 because of incremental snapshot?

Does this mean that deleting older snapshots to manage storage can compromise the ability to fully restore the index?

---

<div class="post-metadata">

**Author:** ![kitkit](https://avatars.discourse-cdn.com/v4/letter/k/c89c15/32.png) [@kitkit](https://forum.opensearch.org/u/kitkit)\
**Post date:** [August 28, 2024, 8:33am UTC](https://forum.opensearch.org/t/how-snapshot-work-and-maintain/20995/2 "2024-08-28T08:33:08Z")

</div>

Tested from my end

In summary, if I delete all snapshots except the latest one, I can still restore all documents, including the oldest ones, by restoring with the latest snapshot.