I have met some difficulties in creating the cluster with openshift v4.2 on GCP (free-tier account). The problem is related to the amount of resources required to create a cluster.
I have changed "machine_type" in terraform variables (openshift/installer/data/data/gcp/master/variables.tf ) but GCP still gives me error like this :
Error: Error waiting for instance to create: Quota 'CPUS_ALL_REGIONS' exceeded. Limit: 12.0 globally. ERROR Quota 'SSD_TOTAL_GB' exceeded. Limit: 250.0 in region europe-west6.
Is there a way to minimize the resources needed to create the cluster ? (I can't modify the GCP QUOTAS.)
Following the documentation for OpenShift UPI in GCP you can deploy VMs of any size you like.
The smallest Red Hat supported configuration is a 3-master 0-worker cluster.
If you didn't care about ensuring your cluster is in a supported configuration you can deploy as small as a single large master.
The real problem you're encountering though is your GCP limits. You can request to have your limits increased and that will also solve your issue.
If you are just wanting to play around with OpenShift you could try CodeReady Containers instead, which is an OpenShift cluster intended to be installed on a laptop.
I have found the files to modify the machine-types : ./openshift/installer/pkg/asset/machines/gcp/machines.go
I have changed the machine_types from n1-standard-4 to n1-standard-1 , Also I have changed pd-SSD to pd-standard , It helps to bypass global free-tier quotas on cpu limits and ssd limits !
Also I have found 1 more problem in timeouts. Because of my machines are weak, the default 30-min timeouts I have set to 180 minutes, cause it takes more time to create the cluster.
I hope it will help someone.
Unfortunatelly I'm still trying to create the cluster, and I will write back if it will be created successfully.
Bootstrap journalctl -b -f -u bootkube.service gives me this PICTURE OF ERROR
