- 12 5月, 2020 10 次提交
-
-
由 Anders F Björklund 提交于
-
由 Medya Ghazizadeh 提交于
docker/podman solution message for non-amd64 archs
-
由 Medya Gh 提交于
-
由 Medya Gh 提交于
-
由 Medya Gh 提交于
-
由 Medya Ghazizadeh 提交于
Don't run the extraction tar container for podman
-
由 Medya Ghazizadeh 提交于
podman: Wrap the start command with cgroup manager too
-
由 Medya Ghazizadeh 提交于
The podman cp command is not in podman-remote
-
由 Medya Ghazizadeh 提交于
Add a PointToHost function for podman as well
-
由 Thomas Strömberg 提交于
Run 'minikube logs' when sub-tests fail, remove some extraneous logging
-
- 11 5月, 2020 5 次提交
-
-
由 Anders F Björklund 提交于
-
由 Medya Ghazizadeh 提交于
Docs: podman-remote is also called podman
-
由 Anders F Björklund 提交于
The "podman-remote" program is installed as "podman" on platforms without a local client (i.e. not Linux) So we need different documentation for different OS. Eventually it will be called "podman" everywhere...
-
由 Anders F Björklund 提交于
And add an active variable, similiar to docker
-
由 Anders F Björklund 提交于
So use podman exec stdin/stdout as a workaround.
-
- 10 5月, 2020 2 次提交
-
-
由 Medya Ghazizadeh 提交于
Podman: disable selinux labels when extracting the tarball (permissions error)
-
由 Anders F Björklund 提交于
The preloaded images will be extracted anyway, using ssh. But if creating them on the volume before the container is booted, means that /var will not be fully copied over to it. And without /var/lib/dpkg and others, kicbase will not boot. So skip the parallel extraction for podman, do it afterwards. Probably shouldn't mount all of /var, but just a sub-set...
-
- 09 5月, 2020 23 次提交
-
-
由 Anders F Björklund 提交于
-
由 Anders F Björklund 提交于
-
由 Anders F Björklund 提交于
Since it doesn't work for either "run" or "start" Error: unknown flag: --cgroup-manager
-
由 Anders F Björklund 提交于
When running with podman, to match the run cmd
-
由 Medya Gh 提交于
-
由 Medya Gh 提交于
-
由 Thomas Stromberg 提交于
-
由 Thomas Stromberg 提交于
-
由 Thomas Stromberg 提交于
-
由 Medya Ghazizadeh 提交于
Wait to add aliases to /etc/hosts before starting kubelet
-
由 Medya Ghazizadeh 提交于
Document that minikube with kubeadm needs 2 CPUs
-
由 Priya Wadhwa 提交于
-
由 Priya Wadhwa 提交于
-
由 Priya Wadhwa 提交于
-
由 Priya Wadhwa 提交于
-
由 Thomas Strömberg 提交于
TestFunctional/parallel/ServiceCmd: Add logging
-
由 Thomas Stromberg 提交于
-
由 Priya Wadhwa 提交于
-
-
由 Priya Wadhwa 提交于
-
由 Priya Wadhwa 提交于
This way, we won't ever enable kubelet to start when systemd starts, preventing a race condition with `/etc/hosts`
-
由 Priya Wadhwa 提交于
This makes sure the kubelet picks up the correct IP
-
由 Priya Wadhwa 提交于
kubelet manages the value of `/etc/hosts` within running containers (e.g. kube-controller-manager). It takes this file from the host, but this causes a race condition between: 1. Us updating /etc/hosts with aliases on IP changes 1. kubelet starting k8s with current value of /etc/hosts This change removes automatically starting kubelet when systemd comes up. Instead, we start it immediately after making our updates to `/etc/hosts`
-