1.1.1 FAQs for Instances
Why Can't I Select Two AZs?
To improve the fault tolerance and HA of a Kafka instance, you are advised to select three AZs when creating the instance. Distributing replicas across three AZs provides a higher level of fault tolerance and availability. This way, if an AZ fails, the system can still work properly. This ensures data reliability and availability in case of a fault of the Kafka cluster.
Why Can't I View the Subnet and Security Group Information When Creating an Instance?
If you cannot view VPC, subnet, security group, and elastic IP during instance creation, you may have not created the related network instance. Please purchase the corresponding network instance.
How Do I Select Storage Space for an Instance?
Disk size = Average traffic × Message retention period × 3 (replicas). We recommend you optimize the topics when migrating to the cloud to reduce costs.
How Do I Choose Between High I/O and Ultra-high I/O?
The choice of Kafka disk type mainly depends on your demands and budgets. Featuring faster random read and write and lower access latency, high I/O disks provide better performance and throughput For more information, see the section of How Do I Choose Between High I/O and Ultra-high I/O.
What Are the Supported Kafka Broker Versions?
Distributed Message Service (Kafka) supports broker versions 2.13-2.8.2.
Can I Modify the Port for Accessing a Kafka Instance?
No. The major considerations are as follows:
l Stability and security: Kafka must work highly reliably and stably. Allowing dynamic changes to access ports can introduce unnecessary risks and instability that could lead to service interruptions or security issues.
l Communication mechanism Kafka works on a distributed cluster architecture, where nodes need to communicate with one another and exchange data over specific ports. Changing the port during runtime may interrupt communication or cause inconsistent communication between nodes.
l Consistent configuration: The nodes in a Kafka cluster must have consistent configurations, including port settings. Modifying the port during running may cause configuration inconsistency and affect the stability of the entire cluster.
Will a Kafka Instance Be Restarted After Its Enterprise Project Is Modified?
No.
Can I Delete a Periodical Kafka Instance?
Distributed Message Service (Kafka) provides comprehensive and thoughtful services to support users' unsubscription needs. You can delete an instance if it is no longer needed.
Can I Change an Instance from Single-AZ Deployment to Multi-AZ Deployment?
No. The AZ configuration cannot be changed once the instance is purchased. To use multiple AZs, please buy another instance.
Does Kafka Support Cross-AZ Disaster Recovery? Where Can I View the AZs Configured for an Existing Instance?
Kafka supports cross-AZ disaster recovery. Depending on the reliability requirements of your data and services, you have a variety of options. You can deploy Kafka instances in one AZ (a single server room) or across multiple AZs (intra-city disaster recovery AZs). For details, see section 8.1.3 Does Kafka Support Multiple AZs? For a purchased instance, you can view azName in Brokers details on Main Menu - Cluster Information.
Does Kafka Support Disk Encryption?
No.
Can I Change the VPC and Subnet After a Kafka Instance Is Created?
No. Once an instance is created, its VPC and subnet cannot be changed.
Can I Upgrade Kafka Instances?
No. Kafka instances cannot be upgraded once they are created.
How Do I Bind an EIP Again?
Log in to the console. On the instance list page, click Set Public IP in the row of the target instance, and click Bind. In the pop-up window, select the EIP to be bound, and click OK.
Note:
When binding a public IP to Kafka, take appropriate security measures. You can use security groups or user permissions to restrict access to protect the security of the Kafka cluster.
Is Data Migrated During Capacity Expansion?
No. Data is not migrated during capacity expansion.
1.1.2 Are Kafka Instances Created in Cluster Mode?
Yes. Kafka instances are generally created in cluster mode. A Kafka cluster consists of multiple nodes and each node stores and processes messages. The cluster mode supports HA, fault tolerance, and scalability.
In a Kafka cluster, each node can work as a leader or replica in multiple partitions, achieving redundancy and failure recovery. Furthermore, Kafka uses a replication mechanism for HA and fault tolerance, ensuring data reliability and availability in the event of node failures.
Note:
You can determine the scale and configuration of a Kafka cluster based on your application requirements and available resources. Specifically, you can configure the number of nodes and partitions and the replication strategy of data in the cluster to cater to your requirements of performance, availability and scalability.