<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
Feilong,
<div class=""><br class="">
</div>
<div class="">Great to see all of these improvements getting into the release.</div>
<div class=""><br class="">
</div>
<div class="">Would it be possible to add these to the release highlights documentation at <a href="https://releases.openstack.org/train/highlights.html" class="">https://releases.openstack.org/train/highlights.html</a> ?</div>
<div class=""><br class="">
</div>
<div class="">Tim<br class="">
<div><br class="">
<blockquote type="cite" class="">
<div class="">On 24 Oct 2019, at 07:56, feilong <<a href="mailto:feilong@catalyst.net.nz" class="">feilong@catalyst.net.nz</a>> wrote:</div>
<br class="Apple-interchange-newline">
<div class="">
<div class="">Hi all,<br class="">
<br class="">
Glad to see the Train just released. I'd like to take this opportunity<br class="">
to highlight some cool features Magnum team have done in this cycle.<br class="">
Please feel free to pop up at #openstack-containers for any feedback.<br class="">
Thank you.<br class="">
<br class="">
1. Fedora CoreOS driver<br class="">
<br class="">
Finally we get this done in this cycle which is quite important and<br class="">
urgent. Because the Fedora Atomic will be end of life soon (at the end<br class="">
of this Nov?). Now with stable/train, user can create k8s cluster based<br class="">
on Fedora CoreOS image. We're welcome for any feedback about this.<br class="">
<br class="">
2. Node group<br class="">
<br class="">
This is a quite large feature, thanks for the great work from CERN team<br class="">
and StackHPC. With this feature, user can create different node groups<br class="">
with different specs.<br class="">
<br class="">
3. Rolling upgrade<br class="">
<br class="">
Now user can do rolling upgrade for both k8s version and the node<br class="">
operating system without downtime, it's a quite important feature.<br class="">
<br class="">
4. Auto healing<br class="">
<br class="">
With auto healing feature, a small service will be deployed on k8s<br class="">
cluster to monitor the health status of the cluster and replace broken<br class="">
node when failure detected.<br class="">
<br class="">
5. Volume based k8s nodes<br class="">
<br class="">
Now user can create a Kubernetes cluster with nodes based on volume and<br class="">
the volume type is configurable. User can even set the volume type of<br class="">
etcd volume which is very useful for cloud providers who want to use SSD<br class="">
for etcd and k8s nodes.<br class="">
<br class="">
6. Private cluster<br class="">
<br class="">
As a security best practice, isolating Kubernetes clusters from Internet<br class="">
access is one of the most desired features for enterprise users. Now<br class="">
user has the flexibility to create a private cluster by default, only<br class="">
allow the API exposed on Internet or fully opened.<br class="">
<br class="">
-- <br class="">
Cheers & Best regards,<br class="">
Feilong Wang (王飞龙)<br class="">
------------------------------------------------------<br class="">
Senior Cloud Software Engineer<br class="">
Tel: +64-48032246<br class="">
<a href="mailto:flwang@catalyst.net.nz" class="">Email: flwang@catalyst.net.nz</a><br class="">
Catalyst IT Limited<br class="">
Level 6, Catalyst House, 150 Willis Street, Wellington<br class="">
------------------------------------------------------ <br class="">
<br class="">
<br class="">
</div>
</div>
</blockquote>
</div>
<br class="">
</div>
</body>
</html>