Data replication

更新时间:
复制 MD 格式

Use the data replication feature of Object Storage Service (OSS) to automatically copy objects from one bucket to another. The source and destination buckets can be in the same or different regions and can belong to the same or different accounts.

What gets replicated

OSS replicates the following data:

  • All objects — both historical objects that existed before the rule takes effect and new or updated objects added afterward

  • Objects encrypted by using Key Management Service (KMS)

  • Object metadata

  • Object ACLs

  • Object tags

  • Multipart upload parts and the combined objects assembled from them

What is not replicated

The following data is excluded from replication:

  • Storage class of objects

  • Last access time of objects (x-oss-last-access-time)

  • Object replicas already created by another replication rule — replication chains are not supported.

    For example, if you configure rules A→B and B→C, objects replicated to B are not further replicated to C.

  • Bucket-level configurations

    Including lifecycle rules, cross-origin resource sharing (CORS) settings, static website hosting settings, hotlink protection settings, and bucket policies

When to use cross-region replication

Use cross-region replication (CRR) when your scenario requires data to exist in a different geographic region:

  • Compliance

    OSS enables data redundancy by default, but if regulations require data copies in geographically dispersed data centers, CRR helps you meet those requirements.

  • Reduced latency

    Keep replicas in OSS data centers closer to your users to minimize data access latency.

  • Disaster recovery

    Maintain a complete copy of all data in a separate region so that if a regional disruption occurs, the replica in the secondary location keeps your application running.

  • Data migration

    Move data from one data center to another with continuous synchronization during the transition period.

  • Multi-region compute

    If compute clusters in two regions analyze the same dataset, use CRR to keep object replicas available in both regions.

When to use same-region replication

Use same-region replication (SRR) when your scenario requires data copies within the same region:

  • Compliance

    If regulations require data to remain within a specific country or region, replicate from the source bucket to multiple destination buckets in the same geographic region.

  • Test and production sync

    Keep your test environment in sync with production without copying data across regions, so you can validate new features against real data without affecting production integrity.

  • Log aggregation

    Consolidate logs stored across multiple buckets or accounts into a single bucket for centralized processing.

References