View a markdown version of this page

Playbook: Amazon OpenSearch Service to Amazon OpenSearch Serverless NextGen - Migration Assistant for Amazon OpenSearch Service

Playbook: Amazon OpenSearch Service to Amazon OpenSearch Serverless NextGen

This playbook is an end-to-end runbook for migrating an existing Amazon OpenSearch Service domain to an Amazon OpenSearch Serverless NextGen collection, using a vector-search workload as the worked example. For a small domain — a handful of vector indexes totaling a few tens of gigabytes — plan for roughly 90 minutes end to end, most of which is unattended snapshot and backfill time.

This is a backfill-style migration: Reindex-from-Snapshot (RFS) takes a point-in-time snapshot of the source domain, migrates metadata, and bulk-indexes documents into the collection. Amazon OpenSearch Serverless NextGen does not accept snapshot restores, ISM/ILM policies, ingest pipelines, custom plugins, or _cluster/settings changes, so the runbook below builds the collection’s shape with security policies and metadata migration rather than a restore. For the broader conceptual comparison, see Migrate to Amazon OpenSearch Serverless NextGen.

For an at-a-glance summary of how Amazon OpenSearch Serverless NextGen constrains a migration, read the How Serverless differs callout before you begin.

Note

Amazon OpenSearch Serverless NextGen authenticates every request with AWS Signature Version 4 (SigV4) — there is no basic-auth or fine-grained access control (FGAC) username and password on the target. Both the snapshot source access and the migration’s write access to the collection are governed by AWS Identity and Access Management (IAM) and by the collection’s data access policy.