AWS solutions documentation change
Summary
Updated instructions for applying configuration changes in workflow-managed runs. Replaced manual tool stop commands with workflow edit/resubmit process and clarified destructive index clearing.
Security assessment
The changes focus on operational procedures for configuration updates and workflow management. No security vulnerabilities or security features are mentioned. The --acknowledge-risk flag exists for operational awareness, not as a security fix.
Diff
diff --git a/solutions/latest/migration-assistant-for-amazon-opensearch-service/ttm-restart-required.md b/solutions/latest/migration-assistant-for-amazon-opensearch-service/ttm-restart-required.md index 1be930ad1..ba02353c8 100644 --- a//solutions/latest/migration-assistant-for-amazon-opensearch-service/ttm-restart-required.md +++ b//solutions/latest/migration-assistant-for-amazon-opensearch-service/ttm-restart-required.md @@ -11 +11 @@ -Whenever you change the type-mapping transformer configuration — including the `multiTypeBehavior` value or the `TypeMappingSanitizationTransformerProvider` settings — you must stop and restart the backfill and replay tools so they pick up the new behavior. Previously migrated data and metadata may also need to be cleared first to avoid an inconsistent state, because indexes and documents that were already written using the old configuration will not be rewritten automatically. +Whenever you change the type-mapping transformer configuration for a workflow-managed run, edit the workflow configuration and resubmit the workflow so metadata migration, backfill, and replay use the same behavior. Previously migrated data and metadata may also need to be cleared first to avoid an inconsistent state, because indexes and documents that were already written using the old configuration will not be rewritten automatically. @@ -13 +13 @@ Whenever you change the type-mapping transformer configuration — including the -To apply a configuration change cleanly, stop the running tools, clear any affected data on the target, and resubmit: +To apply a configuration change cleanly, edit the workflow configuration, clear any affected data on the target when no workflow pod is writing to it, and resubmit: @@ -16,3 +15,0 @@ To apply a configuration change cleanly, stop the running tools, clear any affec - console backfill stop - console replay stop - console clusters clear-indices --cluster target --acknowledge-risk @@ -19,0 +17 @@ To apply a configuration change cleanly, stop the running tools, clear any affec + console clusters clear-indices --cluster target --acknowledge-risk @@ -21,0 +20,2 @@ To apply a configuration change cleanly, stop the running tools, clear any affec +`workflow submit` stops and replaces an existing workflow with the same name before submitting the next run, but it does not clear indexes or documents that were already written to the target. If an abandoned run left migration custom resources that block resubmission, use `workflow reset` to remove those resources before you resubmit. + @@ -24 +24 @@ To apply a configuration change cleanly, stop the running tools, clear any affec -`console clusters clear-indices --cluster target --acknowledge-risk` is destructive — it deletes all indexes on the named cluster. Run it only against the Amazon OpenSearch Service domain or Amazon OpenSearch Serverless collection you are migrating into, and only when you intend to re-run the migration from a clean state. +`console clusters clear-indices --cluster target --acknowledge-risk` is destructive — it deletes all indexes on the named cluster. Run it only against the Amazon OpenSearch Service domain or Amazon OpenSearch Serverless NextGen collection you are migrating into, and only when you intend to re-run the migration from a clean state.