Determine Supplemental Installations
The basic Workload Manager installation initiated through the Suite Admin provides all of the functionality needed to configure your environment and let users model, deploy, and manage applications. However, this basic installation is sufficient only in certain environments as summarized in the Architecture section. Specifically, the following conditions must be true for the basic installation to be sufficient:
All VM-based cloud regions can access the Cisco-hosted bundle store and package store.
The CloudCenter Suite cluster can initiate communication with the cloud API endpoints for all target cloud regions.
The Workload Manager agent running on VMs in all of the VM-based target cloud regions can initiate communications with the CloudCenter Suite cluster.
All VM-based cloud regions support dynamic bootstrapping: the ability to inject a bootstrap script to a VM dynamically upon launch of that VM.
All target Kubernetes clouds have access to the public Docker registry at docker.io.
If any of these conditions are not true, you will need to perform the supplemental installations per the following table.
|Condition motivating supplemental installation||Supplemental component||Where deployed or installed||How Obtained||Install Instructions|
The Cisco-hosted package store or bundle store are not accessible from your VM-based target cloud region
You want to reduce latency associated with downloading artifacts from package store or bundle store
For AWS, AzureRM, and OpenStack: Local Repo Appliance
For Google: A Linux VM with the bundle store and package store manually installed by you
Deployed as a VM in the target VM-based cloud region.
Local Repo Appliance: Obtain the appliance image as explained in Conditional Component Appliance Images
Bundle store and package store installers: Download the appropriate installer script contained in the artifacts.zip file at software.cisco.com
For Local Repo Appliance: Local Repo Appliance (Conditional)
For installing the bundle store manually: Local Bundle Store (Conditional)
For installing the package store manually: Local Package Store (Conditional)
The CloudCenter Suite cluster cannot initiate communications with a cloud region API endpoint
Workload Manager agents in your VM-based cloud region cannot initiate communications with the CloudCenter Suite cluster
For VM-based cloud regions: Deployed as one or more VMs in the target cloud.
For Kubernetes clouds: Deployed as one or more VMs accessible from the target Kubernetes cloud.
|Obtain the appliance image as explained in Conditional Component Appliance Images||See Cloud Remote (Conditional)|
Your VM-based cloud region does not support dynamic bootstrapping
You want to deploy a VM based on your own custom logical image
For CentOS 6: CentOS 6 Worker Image
For all other base OSes: A physical image create by you using the Linux or Windows agent installer script
|Imported to or shared with your VM-based cloud region|
For CentOS 6: Obtain the worker image as explained in Conditional Component Appliance Images
For all other base OSes: Download the appropriate installer script contained in the artifacts.zip file at software.cisco.com
|For creating your own worker image: see Management Agent|
|Your target Kubernetes cloud cannot initiate connections to the internet||Local Docker Registry containing the public Docker images used by your container-based application tiers||Installed on a VM accessible from the target Kubernetes cluster.|
The registry is part of Docker.
The Docker images are available from Docker Hub
Module Update Considerations
When updating the Workload Manager module, be aware that the update occurs over the course of several minutes. During that time there may be a loss of connectivity between the CloudCenter Suite and individual cloud regions even after the Suite Admin UI indicates that the update has completed. Therefore users are encouraged to keep this potential loss of connectivity in mind before applying WM updates.
- No labels