Data Disaster Recovery

MinIO supports the establishment of a disaster recovery center through remote data backup or active-active deployment, ensuring that existing data will not be lost or damaged in the event of a disaster, and ensuring the security and reliability of the data.

Applicable scenarios

• Hot Standby: There are two data centers in the same city or different locations, one serving as the primary and the other as the standby. Data is continuously replicated from the primary cluster to the standby cluster to ensure data consistency. In the event of a disaster in the primary cluster, business traffic can be seamlessly switched to the standby cluster to ensure business continuity.

• Dual Active Data Centers: In a dual active (multi-cluster) architecture, there are two data centers located in different clusters. Both data centers are active and capable of receiving business traffic simultaneously. When one data center experiences a disaster, the business can continue to run uninterrupted in the other data center.

Explanation of nouns

• Primary Cluster: Refers to the cluster that is currently active and processing business requests. It is the source of data or the initiator of operations. In the primary cluster, data is created, modified, or updated, and business traffic is first sent to this cluster for processing.

• Target Cluster: Refers to the cluster that receives data replication, migration, or failover. It is usually in a backup or standby state, waiting to obtain data from the primary cluster or take over business traffic. When the primary cluster fails or needs to be switched, the target cluster receives data replicas from the primary cluster or takes over business traffic to ensure business continuity. In an active-active scenario, the two clusters can serve as target clusters for each other.

Prerequisites

• Both the source cluster and the target cluster need to enable external network access. Please refer to Creating Object Storage for specific configuration methods.

• The primary cluster must use the LoadBalancer access mode, and the target cluster is recommended to support load balancing functionality.

• The source cluster and the target cluster must use the same access protocol, either both using HTTP or both using HTTPS.

• When using the HTTPS protocol, both the primary cluster and the target cluster need to configure their own and each other’s DNS domain resolution, and the two clusters need to share the same set of certificates.

Procedure of operation

1. On the left navigation bar, click Storage > Object Storage.

2. Click on Add Target Cluster under the Data Disaster Recovery tab.

3. Configure the relevant parameters of the target cluster according to the following instructions.

   Parameter	Description
   Access Address	The public access address of the target cluster, starting with http:// or https://.
   Access Key	The access key ID of the target cluster. It is a unique identifier associated with the private access key. The access key ID is used together with the private access key to encrypt and sign requests.
   Secret Key	The private access key used in conjunction with the access key ID to encrypt and sign requests. It identifies the sender and prevents requests from being modified.

4. Click Add.

   • After successful addition, you can view the status of the target cluster and the synchronization status between clusters.

     Parameter	Description
     Cluster Status	The status of the target cluster, including Healthy, Abnormal, or Unknown.
     Storage Buckets	The number of storage buckets to be synchronized or already synchronized. In the hot backup scenario, the number of storage buckets to be synchronized is the number of buckets that the primary cluster needs to synchronize to the target cluster. In the active-active scenario, the number of storage buckets to be synchronized is the sum of the buckets that both the primary cluster and the target cluster need to synchronize with each other.
     Objects	The number of objects that failed to synchronize in the storage bucket. Note: During the MinIO synchronization process, the configuration files related to itself are also synchronized, so this number is for reference only.
     Network Traffic Rate	The network inbound and outbound rate of the primary cluster. In the hot backup scenario, the network inbound rate is always 0. In the active-active scenario, both the inbound and outbound rates have data.

   • If the target cluster fails to be added, you can click Re-add to clean up the cluster information and return to the page for adding the target cluster. You can then re-add the target cluster.

Relevant operations

When data disaster recovery is no longer needed, click Remove target cluster. Removing the target cluster will not delete the synchronized data; if data is currently being synchronized, it will be interrupted.