I figured out the problem, flanneld was not running on the master node.
Cheers,
Rares
On Monday, May 21, 2018 at 12:50:31 PM UTC-7, Rares Vernica wrote:
>
> Hello,
>
> I have a nine bare metal hosts setup with CentOS Atomic. I followed their
> guide on how to setup Kubernetes from
> http://www.p
Thanks Tim!
On Thu, May 24, 2018, 11:55 'Tim Hockin' via Kubernetes user discussion and
Q&A wrote:
> Yes, but any given GCP project can have multiple clusters, but the LB
> names are flat across them. We need to make sure LB names don't collide.
> That's why we use the UUID - the name is not ex
Yes, but any given GCP project can have multiple clusters, but the LB names
are flat across them. We need to make sure LB names don't collide. That's
why we use the UUID - the name is not exactly random - i is a UUID that
maps back to the kubernetes Service.
On Wed, May 23, 2018 at 5:38 PM Jonat
Hello there,
Seems we are facing a quite weird issue where TCP connections are getting
closed forcefully.
How have we seen this? Our scripts speak to a service that is an API. From time
to time, there are connections that scripts see this error: ConnectionError
'Remote end closed without respo
Hi all,
I'm looking at replacing our current job management system with Kubernetes, and
one big question is about managing HTTP routes to individual applications.
Today we use a templated HAProxy config with a comprehensive set of tests to
verify that any route we care about ends up at the expe
