Retry getting AWS instance (kubermatic#887)

If machine-controller tries to get an instance before it's fully-created,
but after the instance request has been issued, it could
happen that it detects that there's no instance and create it again.
That could result in two or more instances created for one Machine object.
This happens more often in some AWS regions because some regions have
slower instance creation (e.g. us-east-1 and us-west-2).

Signed-off-by: Marko Mudrinić <[email protected]>

Co-authored-by: Marko Mudrinić <[email protected]>

use quay.io/kubermatic/kubelet (kubermatic#885)

Return back --network-plugin=cni (kubermatic#883)

Signed-off-by: Artiom Diomin <[email protected]>

Fix containerd systemd service dependency (kubermatic#882)

Signed-off-by: Artiom Diomin <[email protected]>

Fix containerd isntall on ubuntu (kubermatic#881)

* Add unit-test

Signed-off-by: Artiom Diomin <[email protected]>

Containerd support (kubermatic#871)

* Containerd support

Optional containerd support, with default to docker

Signed-off-by: Artiom Diomin <[email protected]>

* Update fixtures

Signed-off-by: Artiom Diomin <[email protected]>

version 1.23.1

Add optional use-octavia flag to openstack cloud config (kubermatic#870)

Include Kubernetes 1.20 in e2e tests (kubermatic#875)

* include kubernetes 1.20 in e2e tests

* update fixtures to the supported versions

* do not use 1.19.6 as there is no kubelet docker image available

fix openstack instanceReadyPeriod/Timeout (kubermatic#848)

Signed-off-by: Moath Qasim <[email protected]>

Signed-off-by: Moath Qasim <[email protected]>