Plan availability zones (AZs)

Availability zones (AZs) are logical groups of resources available for a service. A system administrator can manage the resources available for organizations by assigning them to different AZs. The system supports the following types of AZs:

• Container AZ—A group of worker nodes that provides the same container services. The nodes can be physical servers or virtual machines (VMs) that deploy container applications or services. When you deploy a container image, you select a container AZ. Then, the system automatically deploys that image on a node in the container AZ.

• Compute AZ—A group of virtualized physical servers (hosts) that operate in redundancy mode to provide nonstop compute services. VMs can migrate between hosts in a compute AZ when a host failure occurs. The system offers enhanced compute AZs based on OpenStack Nova compute services.

• Storage AZ—A group of storage spaces for virtual cloud disk services. When you create a cloud disk, you must select a storage AZ. The system offers enhanced storage AZs based on OpenStack Cinder.

Restrictions and guidelines

Developed based on OpenStack Nova, the compute AZs in the system are also called Nova AZs. If you use a Nova AZ for different compute services, you can assign different compute AZ alias to that Nova AZ to identify those compute services.

Developed based on OpenStack Cinder, the storage AZs in the system are also called Cinder AZs.

In the system, each compute node provides a Cinder AZ. When you create a compute node, you must assign it a Cinder AZ name. You can then use a Cinder AZ to provide one or multiple storage services.

Make sure two hosts are available for cloud host migration.

Plan AZs

Follow these rules when planning AZs:

• Plan compute AZs before planning storage AZs.

• Plan different compute AZs for different virtualization platforms, because they provide different features and use different images.

• Assign different compute resource pools to different compute AZs because of limits of the management capabilities of virtualization platforms.

• In the system, each compute node provides a Cinder AZ. When you create a compute node, you must assign it a Cinder AZ name.

	·          When the system is interconnected with a CAS CVM, a compute resource pool in the system is a CAS host pool or a cluster in a host pool. ·          When the system is interconnected with a VMware virtualization platform, a compute resource pool in the system is a VMware cluster.

 

Example

Assume that Company A has the following compute resources:

• The compute resources are managed by two virtualization platforms CAS and VMware. The CAS virtualization platform contains two host pools HPool1 and HPool2, each of which contains 16 CVK hosts. The VMware virtualization platform contains one cluster Cluster1, which contains 16 ESXi hosts.

• Four departments Org1, Org2, Org3, and Org4 use the compute resources in HPool1, HPool2, Cluster1, and Cluster1, respectively.

Table-1 shows the AZ plan of the company.

Table-1 AZ planning example

Dept.	Org1	Org2	Org3	Org4
Hypervisor	CAS		VMware
Hosts	16 CVK hosts	16 CVK hosts	16 ESXi hosts
Compute resource pool	HPool1	HPool2	Cluster1
Storage AZ	Cinder1	Cinder2	Cinder3
Compute AZ	Nova1	Nova2	Nova3
Storage AZ alias	Org1 AZ	Org2 AZ	Org3 AZ	Org4 AZ
Compute AZ alias	Org1 AZ	Org2 AZ	Org3 AZ	Org4 AZ