Basic Introduction

Data replication is the feature of automatically and asynchronously copying files (objects) between storage spaces (buckets) in BOS Data Center (Region). It will copy the creation, update and deletion of objects from the source storage space to the target storage space.

The data replication feature can provide bucket disaster recovery across regions or meet the needs of user data replication. The objects in the target bucket are exact copies of the objects in the source bucket with the same object name, metadata, and content, such as creation time, owner, user-defined metadata, object ACL, object content, and so on.

Usage Scenarios

You may configure data replication for a bucket for a variety of reasons, including:

• Efficient access and reduced latency: When accessing bucket or object across regions, the access speed will be reduced due to geographical differences. Through the data replication feature, the required data can be synchronized to the target region in advance, thus improving the access efficiency.
• Cross-region disaster recovery: It is required to be based on the compliance requirements of finance, government and other industries, although there are multiple copies of data in the same Region. However, in order to prevent computer room-level disasters such as floods and earthquakes, it is also required to keep a copy of data across Region. BOS’s data synchronization feature provides such capability.
• Cross-account copy: Applicable to multiple account scenarios for enterprises. In consideration of data security and multiple backups, enterprises hope to synchronize the data from account A to account B regularly, so as to realize the cross-account copy and sharing of data.
• Cross-regional data reuse: Due to business reasons, such as the need to use the same set of data for computing clusters built across regions, BOS data replication feature can be used to copy data copies.

Operation Method

BOS supports both the use of console and the use of SDK for cross-regional replication, as follows:

• Use Console to Conduct Data Replication

Use API for data replication

• PutbucketReplication
• GetbucketReplication
• DeleteBucketReplication
• GetbucketReplicationProgress

Use SDK for data replication PHP SDK

Instructions for Use

• You can specify the data to be synchronized in the source bucket by defining the prefix of the file name, or you can directly synchronize all the data in the source bucket.
• The storage type of the target bucket can be selected to be the same as that of the source object, or other storage types can be selected as needed.
• To copy the data of non-standard Object, the retrieval charge will be triggered.
• The archive storage is supported for the storage type of the target Bucket files. After the rules are successfully added, it is available to see, edit and delete the already synchronized strategy in the current Bucket .
• After the historical file copy is enabled, the historical Object will match the prefix rules in the incremental data, and only the files sharing the same prefix will be synchronized; if you specify to synchronize all data in the Bucket, the history will also be synchronized.
• Two Buckets of data synchronization can be cross-region or in the same region. Only data synchronization across regional Buckets will trigger traffic charges, and data synchronization with regional Buckets will not charge traffic charges.
• The two Buckets with synchronized data can be from cross-account or same account. Please make sure that the target Bucket name you entered are accurate and correct. If the target Bucket in the target account is located in other area, the cross-region traffic costs will be incurred, and the traffic costs will be borne by the source account, and the target account shall assume the storage and request costs.
• Support for synchronization of multiple rules among Buckets, that is, a source Bucket can have multiple target Buckets, and a target Bucket can have multiple source Buckets. Suppose there are three Buckets named A, B, and C.

• • Support for using A as the source Bucket of B while using B as the source Bucket of C;

    • Support for using A as the source Bucket of B while using B as the source Bucket of A;
    • Support for using A as the source Bucket of B and C at the same time;
    • Support for using C as the source Bucket of A and B at the same time.
• For the sake of data security, BOS will not access and return the enabled status of cross-account synchronization. If you have set up and enabled the cross-account synchronization, please query proactively whether the data synchronization in the account has been enabled. If the source data have not been synchronized to the Bucket of target account within 5 minutes, please check again whether Write Privilege has been enabled. If the write permission has been enabled but the synchronization fails, please submit a Ticket.

Restriction Description

• For two buckets that are synchronized, since you can operate both buckets at the same time, the copy of the past object by the source bucket may overwrite the object with the same name in the target bucket, please pay attention when using it.
• Since data replication adopts asynchronous replication, it takes a certain time for data to be replicated to the target bucket, usually ranging from several minutes to several hours, depending on the size of the data.