Document URL: https://docs.openshift.com/container-platform/4.5/installing/installing_vsphere/installing-vsphere-installer-provisioned.html#installation-vsphere-installer-infra-requirements_installing-vsphere-installer-provisioned Section Number and Name: - Cluster resources - Networking requirements Describe the issue: While doing OCP 4.5 installation on vSphere using IPI method, there is confusion happening related to the required VIPs and the description mentioned in the table format in DNS records section. It is not clear that the required VIPs are just two reserved IPs and for those particular VIPs the DNS records are needed. Also, the Description mentioned in the Required DNS records Table has clauses like "load balancer for the control plane machines" and "the load balancer that targets the machines that run the Ingress router pods". If we are not explaining how installer is creating the haproxy and keepalived pods in openshift-vsphere-infra namespace then the understanding of backend of the required VIPs becomes more confusing. So, adding information about how backend is configured for those particular VIPs would be helpful. Also, in Cluster resources section, the minimum storage required for standard cluster is mentioned. But in the same way, the information about the CPU and Memory which will get consume during the default IPI installation would be valuable to analyse the vSphere infrastructure resources prior to installation and avoiding pending pods due to resource unavailability. Suggestions for improvement: I have suggested the improvements along with the problem description. Additional information:
I have a question to help me complete this request: In the Cluster Resources section, you'd like information about the CPU and memory that are consumed during the default IPI installation. I can definitely add this, but I don't know what that information is. Do you have these details, or know who I can reach out to? I've drafted some changes to the DNS Records section, so once I understand the details of what needs to be added for the Cluster Resources section, I can get a PR up for you to review.
You may need to check this with engineering team what are the default resources consumed by the bootstrap, master and compute nodes as they have recommended for storage.
Patrick - I got your name from Cody Hoag as someone who would probably be able to advise on this. Could you please provide some info on the resources consumed by the bootstrap pod during IPI installation on vSphere?
Bootstrap in IPI uses same resources as documented for UPI: vCPU: 4 Virtual RAM: 16 GB Storage: 120 GB https://docs.openshift.com/container-platform/4.6/installing/installing_vsphere/installing-vsphere.html#minimum-resource-requirements_installing-vsphere Bootstrap node resources, unlike masters/compute, are hard-coded and non configurable. Remember that bootstrap is torn down before worker creation, so the max resources used by cluster are greater of (masters + bootstrap or masters + compute). Let me know if you need more info.
Thanks for the quick answer! With that info, the PR is ready for review: https://github.com/openshift/openshift-docs/pull/27842 Aditya - I can't find your GitHub username to tag you, but please take a look. Patrick - tagged you to make sure I didn't mangle it.
PR has rotted without details, please reopen for a current release if still needed.
Reopening the bugzilla, as required information can be added to new OCP versions to make documentation better.
Hi Aditya - did you ever get a chance to review the PR for this doc to see if it met your needs? https://github.com/openshift/openshift-docs/pull/27842 It closed waiting on a few answers but if you think it looks good I will open a fresh new one to get this into 4.6+. Please let me know.
PTAL at https://github.com/openshift/openshift-docs/pull/27842 and let me know if these changes are what you are looking for. If so, I will open a new one for 4.6+
The PR mentioned about the static IP requirement but not where actually these IPs point to in backend of OCP. There should be more explanation added to documentation about haproxy, keepalived pods running in openshift-vsphere-infra namespace which comes into picture when we consider these two static IPs. Additional information would be helpful in understanding.