Description of problem: I configured an OpenShift ansible installation with crio openshift_use_crio: true openshift_crio_enable_docker_gc: true When the installer tried to start the nodes it failed on all on them. This is the node service log: 36:13.866837 33661 remote_runtime.go:43] Connecting to runtime service /var/run/crio/crio.sock 36:13.866861 33661 util_unix.go:75] Using "/var/run/crio/crio.sock" as endpoint is deprecated, please consider using full url format "unix:///var/run/crio/crio.sock". 36:13.866966 33661 util_unix.go:75] Using "/var/run/crio/crio.sock" as endpoint is deprecated, please consider using full url format "unix:///var/run/crio/crio.sock". 36:13.877418 33661 remote_runtime.go:69] Version from runtime service failed: rpc error: code = Unavailable desc = grpc: the connection is unavailable 36:13.877502 33661 kuberuntime_manager.go:172] Get runtime version failed: rpc error: code = Unavailable desc = grpc: the connection is unavailable 36:13.877524 33661 server.go:173] failed to run Kubelet: failed to create kubelet: rpc error: code = Unavailable desc = grpc: the connection is unavailable ce: main process exited, code=exited, status=255/n/a ode. service entered failed state. on firther investigation it looks like crio is not even installed and the crio service is not started: [root@env1-master-m7zc ~]# rpm -qa | grep cri-o [root@env1-master-m7zc ~]# systemctl status crio Unit crio.service could not be found. Version-Release number of selected component (if applicable): atomic-openshift-node-3.9.14-1.git.0.4efa2ca.el7.x86_64 How reproducible: not sure. 100% on my setup. Steps to Reproduce: I can share my inventory if needed. Actual results: node service doesn't start. Expected results: crio service is running, node service can start, containers can be started.
I understood what is going on. In order for the ansible installer to be successful the cri-o rpm must be installed and the crio.service must be started. This bugzilla should be turned into a documentation bug. The host prep section of the documentation should mention to perform the above steps when planning to use crio.
We just published a new CRI-O guide that, I believe, has the proper ansible settings to make an OpenShift installation with CRI-O nodes work properly: https://access.redhat.com/documentation/en-us/openshift_container_platform/3.11/html-single/cri-o_runtime/ The guide also has configuration and troubleshooting sections. If there are issues that we have missed, please open another bug against that document.