AWS solutions documentation change
Summary
Added workflow-managed backup option, instructions for multi-collection backups, S3 path clarification, and schema validation requirements.
Security assessment
Changes focus on backup procedures and data migration workflows. While mentioning S3 buckets and IAM roles, these are operational details rather than security fixes or new security documentation. No vulnerabilities or security enhancements are addressed.
Diff
diff --git a/solutions/latest/migration-assistant-for-amazon-opensearch-service/pb-solr-step1.md b/solutions/latest/migration-assistant-for-amazon-opensearch-service/pb-solr-step1.md index eb2ac343c..c0893e55c 100644 --- a//solutions/latest/migration-assistant-for-amazon-opensearch-service/pb-solr-step1.md +++ b//solutions/latest/migration-assistant-for-amazon-opensearch-service/pb-solr-step1.md @@ -9 +9,5 @@ -First, stop or pause writes to the Apache Solr source so the backup is a consistent point-in-time copy. Then create a backup using the API for your deployment topology, write it to a path that the Migration Assistant can later read, and sync that path to your Amazon S3 migration bucket. +First, stop or pause writes to the Apache Solr source so the backup is a consistent point-in-time copy. This playbook shows the externally managed backup path: you create a backup using the API for your deployment topology, write it to a path that the Migration Assistant can later read, and sync that path to your Amazon S3 migration bucket. + +###### Note + +Migration Assistant can also create the Solr backup during the workflow when the Solr endpoint is reachable from the Migration Console and the Solr S3 backup repository is configured. That path auto-discovers all collections or cores. If you want to use workflow-managed backup creation, skip the manual `curl` commands here and use the `createSnapshotConfig` form in [Step 2](./pb-solr-step2.html). @@ -19,0 +24,2 @@ Wait until the `REQUESTSTATUS` response reports the backup as completed before m +For multiple SolrCloud collections, repeat the backup request for each collection under one shared snapshot root. Use the collection name as the Solr backup `name` and include the shared backup name in the `location`, for example `location=/path/to/backup/<BACKUP_NAME>&name=<COLLECTION>`. After sync, the Amazon S3 path should contain one top-level directory per collection under `<BACKUP_NAME>/`, and the workflow should set `s3RepoPathUri` to the parent path and `externallyManagedSnapshotName` to `<BACKUP_NAME>`. + @@ -33 +39 @@ For SolrCloud, the backup `location` must be a path that is reachable from every -After the backup completes, sync the backup directory to your Amazon S3 migration bucket so RFS can read it: +After the backup completes, sync the parent backup directory to your Amazon S3 migration bucket so RFS can read it. The synced directory should contain the `<BACKUP_NAME>` directory: @@ -38 +44,8 @@ After the backup completes, sync the backup directory to your Amazon S3 migratio -You can use the default Migration Assistant bucket (`s3://migrations-default-<ACCOUNT_ID>-<STAGE>-<REGION>`) or any bucket the migration IAM role can read. Record the resulting Amazon S3 URI — you supply it as `s3RepoPathUri` in the next step. +You can use the default Migration Assistant bucket (`s3://migrations-default-<ACCOUNT_ID>-<STAGE>-<REGION>`) or any bucket the migration IAM role can read. Record the resulting parent Amazon S3 URI — you supply it as `s3RepoPathUri` in the next step. The workflow appends `externallyManagedSnapshotName`, so do not include `<BACKUP_NAME>` in `s3RepoPathUri`. + +Before configuring the workflow, check the S3 layout: + + + aws s3 ls s3://<BUCKET>/solr-backup/<BACKUP_NAME>/ + +The listing should show collection or core directories, not `zk_backup_0/` directly. Inside each collection directory, RFS can read numbered `zk_backup_N/` backups, older bare `zk_backup/` backups, and one-level nested Solr 8/9 incremental layouts. The latest ZooKeeper backup for each collection must include `managed-schema.xml`, `managed-schema`, or `schema.xml` under `configs/<CONFIGSET>/` so metadata migration can translate the schema.