Hardware requirements
You have the option to install IBM Infrastructure Automation (stand-alone) or install IBM Infrastructure Automation with IBM Cloud Pak for AIOps. For more information about requirements, see the following sections.
- Requirements for installing Infrastructure Automation with IBM Cloud Pak for AIOps
- Requirements for installing Infrastructure Automation (stand-alone)
Note: Your hardware must be able to support Red Hat OpenShift, Infrastructure Automation, and your chosen storage solution.
Requirements for installing Infrastructure Automation with IBM Cloud Pak for AIOps
If you are installing Infrastructure Automation with IBM Cloud Pak for AIOps, then you need to follow the hardware requirements for Infrastructure Automation and IBM Cloud Pak for AIOps.
For Cloud Pak for AIOps and Infrastructure Automation requirements, see the following sections:
Requirements for installing Infrastructure Automation (stand-alone)
Use the following sections to review the details about the minimum resource requirements for installing Infrastructure Automation. Use this information for sizing your environment when you are deploying Infrastructure Automation. Multi-region and multi-zone clusters are not supported.
Note: Work with your IBM Sales representative (or Business Partner) to ensure that the allocated hardware resources for your environment are sufficient to best fit your business needs. For requirements for larger production deployments and to make sure that you meet the right level of resiliency, an IBM representative can work with you on your deployment plans. For Infrastructure Automation, you need to provide the following information:
- The number of cloud environments that you connect to.
- The number of OpenStack environments that you manage.
- The average number of VMs in each of the cloud environments.
- The number of on-premises virtualized environments that you manage. Identify whether it is VMware or KVM, and the number of VMs in each virtualized environment.
Resource allocation for the Red Hat OpenShift Container Platform control plane master nodes
The following table details the minimum requirements for the Red Hat OpenShift control plane master nodes that are necessary for an Red Hat OpenShift cluster to withstand the demands of an Infrastructure Automation deployment. Infrastructure
Automation has many Kubernetes operators that interface with the API server and etcd storage, and so master nodes must be adequately sized.
| Node count | vCPU (per node) | Memory (GB) (per node) | vCPU (all nodes) | Memory (GB) (all nodes) |
|---|---|---|---|---|
| 3 | 8 | 16 | 24 | 48 |
Resource allocation for the Red Hat OpenShift Container Platform worker nodes
The following table provides the required Red Hat OpenShift worker resources per node for a sample Infrastructure Automation deployment.
| Node count | vCPU (per node) | Memory (GB) (per node) | vCPU (all nodes) | Memory (GB) (all nodes) | Disk space (GB) | vCPU (extra capacity) | Memory (GB) (extra capacity) |
|---|---|---|---|---|---|---|---|
| 3 | 16 | 32 | 48 | 96 | 212 | 17 | 18 |
| Resource | Deployment (Managed Services and Infrastructure management) No infrastructure management providers (not under load) |
Deployment (Managed Services and Infrastructure management) 1 infrastructure management provider (not under load) |
|---|---|---|
| Worker node count Total vCPUs Total memory (GB) Persistent storage (Gi) |
3 14 35 70 |
3 14 38 70 |
Notes:
- These estimates are for an Infrastructure Automation installation with both Managed services and Infrastructure Management deployed.
- The requirements for Infrastructure Automation are in addition to the requirements for IBM Cloud Pak for AIOps.
- You might need extra worker nodes, resource capacity, or both, depending on your own requirements. The CPU, memory, and disk resources that are required for Infrastructure Automation depend on the number of provider instances and the total
number of instances and VMs managed by the providers. Other worker node counts and sizes can be used based on your available hardware configurations, but the following conditions must be met:
- The combined available vCPU and memory of the worker nodes must equal or exceed the total requirements.
- Each worker node has at least the minimum vCPU and memory that is given in the preceding table.
- Warning: These requirements do not include CPU and memory resources for hosting a storage provider. These storage providers can require more nodes, resources, or both, to run. The extra resources that are needed can vary based on your selected storage provider. A general recommendation is for one extra worker node in a starter deployment, and for three extra worker nodes for a production deployment where HA is required. Consult your storage provider's documentation for exact requirements.