Make sure your hardware requirements reflect the ones reported here: Requirements
CPU Quantity Check
The Head node must have at least 12 CPU cores. Verify that your system has adequate CPU resources.
CPU KVM Check
The nodes must support KVM virtualization on the CPUs being used. You may have to turn it on in the firmware (BIOS, EFI, etc.) of your system. Sometimes a firmware update is required to support this if the system is older.
If you're trying to implement this in a "nested virtualization" environment, the virtualization system must support this and have it turned on for your VM.
Systems that support nested virtualization:
- VMWare - https://communities.vmware.com/docs/DOC-8970 , https://communities.vmware.com/community/vmtn/bestpractices/nested
- Xen - http://wiki.xenproject.org/wiki/Nested_Virtualization_in_Xen
- KVM - https://fedoraproject.org/wiki/How_to_enable_nested_virtualization_in_KVM , https://wiki.archlinux.org/index.php/KVM#Nested_virtualization
- Hyper V - https://msdn.microsoft.com/en-us/virtualization/hyperv_on_windows/user_guide/nesting
Systems that lack nested virtualization:
- Virtualbox - track this feature request: https://www.virtualbox.org/ticket/4032
DNS Lookup Check
The nodes must be able to look up hosts on the internet, to download software, VM images, etc. This verifies that your DNS is working properly using `dig`. if this step fails, try running the playbook again, and if it still doesn't work, check that your DNS settings are working correctly.
DNS Global Root Connectivity Check
A DNS server is run on the head node to resolve internal DNS for the pod, and for non-pod domains it will makes requests to the global DNS root servers. If this step fails, verify that you do not have a firewall that is blocking outgoing DNS requests.
HTTP Download Check
This verifies that you can download files from the internet. If this fails, make sure you don't have a proxy or other system blocking network traffic.
HTTPS Download Check
The same as the HTTP check above, but over HTTPS. If this fails but the HTTP succeeds, check for a proxy, and that the time on the node is accurate.
Check that VM's can be found in DNS
This likely means that something has gone wrong either with the connectivity or setup of the DNS server, or with the client configuration. Verify that
/etc/resolv.conf on all the nodes is using the IP address of the head node's default interface as the DNS server.