Snapshot Storage Gain

Hello, we are setting up one node opensearch cluster, so no replication.But there will be parsing. We are planning to use snapshots to store logs for extended period of time, like several years. Not sure backup storage yet mounted file system or Minio. Is there any difference in used storage between normal indexes and snapshots? like how we can gain in storage? (on node , no replication)

Hi @sukur55 ,

The snapshots are primarily used as backup mechanism in Opensearch. Snapshots could be stored in S3/ACS/GCS/OCI and significantly reduce the costs for data which is not used often (or not changed so you could compromise on replication factor for example), help with disaster recovery and data availability. But you cannot do any operations against snapshots (indexing, searching, aggregations, …) at the moment, for that the snapshot have to be restored into index. Hope it helps.

Best Regards,
Andriy Redko

1 Like

Can I delete fields to save on storage when taking snapshot/backup?

No, at least with default plugins

what about source only snapshots? is it possible?

Source only snapshots are XPack feature of Elasticsearch (at least in 7.10.x where the fork has happened), it is not available in Opensearch.

[1] Register a snapshot repository | Elasticsearch Guide [7.10] | Elastic

yes, you are right. Thank you. Looks like having an active index and fs snapshots almost takes same space(not replicated)