Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: Idbffeba25547f3e1a83d83c86b65ea52cecbc8c5
Issue-ID: OOM-2829
Signed-off-by: Bartek Grzybowski <b.grzybowski@partner.samsung.com>
|
|
Currently used 2.0.0-beta4 supported k8s 1.15 at max
Change-Id: I9565eaa78ebbca48377d65d87c77bc0893b29487
Issue-ID: OOM-2715
Signed-off-by: Bartek Grzybowski <b.grzybowski@partner.samsung.com>
|
|
Manual system_images assignment in rke's cluster config
is dropped as for any particular kubernetes release Rancher
project holds a predefined, static map connecting Rancher's
Kubernetes version with particular system images, so it's
enough to define k8s version.
Ref.: https://rancher.com/docs/rke/latest/en/config-options/system-images/
Change-Id: I027589c86500bbc5260caaa4aef7065d23d2241c
Issue-ID: OOM-2686
Signed-off-by: Bartek Grzybowski <b.grzybowski@partner.samsung.com>
|
|
Kubelet can be configured on startup now - just add new key/values in:
kubelet-static-config.yml.j2
Also kubelet static (startup) configuration can be overriden later
(dynamically) during its runtime - for more info read this:
https://kubernetes.io/docs/tasks/administer-cluster/reconfigure-kubelet/
Issue-ID: OOM-2199
Change-Id: Ia4790a4dfc04f0caed695df7bfd343253bd95bc0
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
|
|
This commit updates rke (implementing K8S 1.15.4), kubectl binary and
helm binary with tiller image:
rke 0.3.0
kubernetes 1.15.4
kubectl 1.15.4
helm 2.14.2
dashboard 2.0.0-beta4
Kubernetes dashboard is also updated to be compliant with the 1.15 K8S.
Note 1:
There is a new datalist to split rke only images from other K8S related
images.
Note 2:
El Alto recommendation cannot be fullfilled to the letter, because rke
does not have any release supporting 1.15.2 version of kubernetes (only
experimental).
Issue-ID: OOM-2122
Change-Id: I02cf0955158eca622bcac40ceed42588dbdd3499
Signed-off-by: Tomáš Levora <t.levora@partner.samsung.com>
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
Signed-off-by: Bartek Grzybowski <b.grzybowski@partner.samsung.com>
|
|
Use the new external_ip inventory variable to populate node address.
Note that cluster_ip is still used if external_ip has not been set.
This makes it possible to connect to kubernetes apiserver using the external_ip
without disabling ssl cert verification.
As an added bonus, the inventory hostname is now used as the node name to make
node names more predictable.
Change-Id: I3cc25c488d68f2f858a5eb3438d96285fa7517ca
Issue-ID: OOM-2112
Signed-off-by: Michal Zegan <m.zegan@samsung.com>
|
|
Plus the option to have etcd on tmpfs (volatile storage) for little
bigger (maybe) speed (NOT FOR PRODUCTION DEPLOYMENT!).
Issue-ID: OOM-2058
Change-Id: I0bbfc4fbae7f6b46e2fae49656437cd6748efd49
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
|
|
Etcd cluster is now placed in members of kubernetes-etcd group instead
of kubernetes-node.
Due to defaults, it means infra is the only etcd
by default.
Change-Id: Iae05a42442849e4a248d73c2d78f5e2b0eae7255
Issue-ID: OOM-1778
Signed-off-by: Michal Zegan <m.zegan@samsung.com>
|
|
Rke playbook/role is modified to take advantage of the new group structure.
Namely, all members of kubernetes group are treated as cluster members
independent of their role.
The role itself is selected in cluster.yml.j2 template
based on actual membership in either kubernetes-node
or kubernetes-control-plane group.
Change-Id: I9a5bbfd090aff17018a610a01d6f88d848fc26c4
Issue-ID: OOM-1778
Signed-off-by: Michal Zegan <m.zegan@samsung.com>
|
|
Either run kubectl proxy, but you must setup routes or kubectl on some
node (only infra has kubectl by default). Another option is connecting
onto the node where the dashboard is exposed via NodePort. Instructions
are here:
https://github.com/kubernetes/dashboard/wiki/Accessing-dashboard
(1.7.X and above)
Change-Id: I5a28f458c0ae949278b1e6eaf6eb063d09f93fc0
Issue-ID: OOM-1778
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
Signed-off-by: Michal Zegan <m.zegan@samsung.com>
|
|
Added a new playbook rke.yml and role rke which uses rancher RKE:
https://github.com/rancher/rke
It's an implementation of the kubernetes from rancher.com folks
and it is an alternative to the official kubernetes tool: kubeadm.
NOTE: Rancher has a notion of a 'control plane' which cannot run with
etcd on all nodes in a multi-node setup. Control-plane node is the
first kubernetes node from the inventory (as of now).
Change-Id: I0bf669442a5183efa20d44fb1cac823e7ce54348
Issue-ID: OOM-1778
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
Signed-off-by: Michal Zegan <m.zegan@samsung.com>
|