[ANN] Canonical Distribution of Kubernetes Release Notes [week of 10/13]

Charles Butler charles.butler at canonical.com
Thu Oct 13 19:53:38 UTC 2016

10/13 Release Notes

Greetings everyone! It's been a short but busy week for us.  We've landed a
lot more bugfixes and some new features for you all to kick the tires.
We're excited to push this week’s rollup as it contains the early work
(alpha?) in consuming Ceph RBD volumes for persistent volume storage in
your kubernetes workloads.

It’s missing from the readme, so here is a quick rundown below the release
notes.  As always, bugs/comments/questions are all welcome.

Or you can find us on irc in #juju on irc.freenode.net



   Added DOCKER_OPTS passthrough config option. This enables end users to
   configure the runtime of their docker-engine (Such as insecure registries)
   without having to add python code to the layers and/or re-build a fork.


   Corrected an immutable config when attempting to switch between archive
   docker package and the docker-engine package from upstream.

Thanks @brianlbaird and @simonklb for driving this feature during



   Corrected the directory glob pattern on flannel which was failing and
   causing some false positives in the cloud weather report testing tool.

Kubernetes Master


   Added a create-rbd-pv action to enlist persistent storage from Ceph.

      This requires the use of the ceph-mon charm from the
      openstack-charmers-next branch.

   Closed a bug where running microbots would yield an EOF error due to
   proxy settings. Consult the README for limited egress environments. (Thanks
   @ryebot and @cynerva)

Kubernetes Worker


   Added a kubectl wrapper for context with manifests, and a kubectl
   wrapper for arbitrary keyword args.

   Various lint fixes.

   Worker nodes now cleanly remove themselves from the cluster during the
   stop hook. (Thanks to @ryebot and @cynerva)

   The ingress controller now scales to the number of deployed worker
   units. 1 ingress controller per 1 worker unit. (Thanks to @ryebot and

Canonical Distribution of Kubernetes Bundle


   Added documentation for proxy settings in network limited environments
   to the bundle README. (Thanks to @ryebot and @cynerva)

   Updated the README with additional limitation notes about which charms
   are not compatible enough to run in LXD at this time.

   Bumped each charm to their latest revision.

Kubernetes Core Bundle

A minimalist bundle, only deploying the bare minimum required to evaluate
kubernetes. Useful when doing laptop development or resource constrained
environments. (Thanks @cynerva and @ryebot)


   The kubernetes core bundle has been updated with our latest release of
   the Canonical Distribution of Kubernetes (CDK) charms.

   Brand new README imported from the CDK bundle.

https://github.com/juju-solutions/bundle-kubernetes-core  - we’re still
testing this minimal bundle, and it will be published in the charm store as
early as next week. Thanks for your early interest!



   Refactored the test to gate on the status messages before reading the
   deployment as ready and proceeding with executing tests.

Quick Rundown on how to enlist RBD PV’s

You’ll need to be running at bare-minimum the ceph-mon charm from the
~openstack-charmers-next namespace.

juju deploy cs:~openstack-charmers-next/xenial/ceph-mon -n 3

juju deploy cs:ceph-osd -n 3

>From here you will need to enlist the OSD storage devices:

For example on GCE you would fulfill this request with GCE Persistent Disks:

juju add-storage ceph-osd/0 osd-devices=gce,10gb

juju add-storage ceph-osd/1 osd-devices=gce,10gb

juju add-storage ceph-osd/2 osd-devices=gce,10gb

This will allocate 30gb of storage, across the 3 OSD device nodes, and
begin your ceph replicated file storage cluster.

Next is to relate the storage cluster with the kubernetes master:

juju add-relation kubernetes-master ceph-mon

We’re now ready to enlist Persistent Volumes in Kubernetes which our
workloads can consume via PersistentVolumeClaims

juju run-action kubernetes-master/0 create-rbd-pv name=test size=50

Tailing a watch on your kubernetes cluster like the following, you should
see the PV become enlisted and be marked as available:

watch kubectl get pv --all-namespaces


test   50M          RWO       Available                              10s

To consume these Persistent Volumes, your pods will need an associated PVC
with them, and is outside the scope of this tutorial. See:
http://kubernetes.io/docs/user-guide/persistent-volumes/  for more

This work is early so please let us know if you using storage and remember
to open issues at:

Juju Charmer
Canonical Group Ltd.
Ubuntu - Linux for human beings | www.ubuntu.com
Juju - The fastest way to model your application | www.jujucharms.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju/attachments/20161013/2d48e609/attachment.html>

More information about the Juju mailing list