I gave that a shot, as well as trying to symlink docker.sock to point to crio.sock — that did not work I’m afraid.. The Kubernetes factory seemed to work just fine, I’m just not clear on what the pros/cons of that are.
On May 14, 2019 at 9:20:59 AM, Don Schenck (don.sche...@gmail.com) wrote: Since Podman is compatible, can you alias it to 'docker' and see if that works? On Tue, May 14, 2019 at 9:14 AM David P Grove <gro...@us.ibm.com> wrote: > Henry Zektser <japha...@gmail.com> wrote on 05/13/2019 08:50:11 AM: > > > > > I guess what I’m trying to articulate is that openwhisk-deploy-kube fails > > in the case of OpenShift 4. I wanted to get the groups thoughts and > > confirm I’m not missing anything before I try to hammer in CRI-O support > — > > which appears like it won’t be that easy. > > Hi Henry, > > It would be great if we could get openwhisk-deploy-kube to work on > OpenShift4. > > I don't have access to an OS4 system to help you directly, but > happy > to provide advice & help get fixes into the code to make it work. > > I'd suggest for the first cut to use the KubernetesContainerFactory > instead of the DockerContainerFactory. This should avoid the invoker > problems. To do this, add the following to your mycluster.yaml for the > Helm template command. > > invoker: > containerFactory: > impl: kubernetes > kubernetes: > agent: > enabled: false > > --dave > -- I'm also on Twitter (@DonSchenck), Facebook (DonSchenck) and Skype (schenckdon).