Hey Oliver!

 

Sorry for seeing this a month later, but awesome job with this blog post.  Are you okay with us linking it in the Cluster API driver documentation as a “reference” blog post?

 

Thanks!

Mohammed

 

From: Oliver Weinmann <oliver.weinmann@me.com>
Date: Saturday, January 6, 2024 at 10:55 AM
To: Franck VEDEL <franck.vedel@univ-grenoble-alpes.fr>, openstack-discuss <openstack-discuss@lists.openstack.org>
Subject: Re: [kolla-ansible][magnum] question

You don't often get email from oliver.weinmann@me.com. Learn why this is important

Hi Franck,

glad that I was able to help. Sorry I don't know about this one. But a quick google revealed this:

https://docs.fedoraproject.org/en-US/fedora-coreos/sysconfig-configure-swaponzram/

Looks as if the FCOS image has this enabled somehow.

I have updated my blogpost on CAPI in MAGNUM. It is finally working and its way faster to deploy Kubernetes. This will no longer use FCOS images but Ubuntu or FLATCAR.

Maybe you want to give it a try:

https://www.roksblog.de/openstack-magnum-cluster-api-driver/

Have a nice weekend.

Cheers,

Oliver

Am 05.01.2024 um 17:07 schrieb Franck VEDEL:

Hi Oliver.

Thanks for you help !!

 

 

 

I understood a lot of things, even though I'm not a Kubernetes expert.

Everything works as I wanted, except for one detail.

Do you know why we get this type of message (and quite a few minutes to wait) when creating a node:

 

zram_generator…………. no configuration found. 

 

This is the last point that I cannot improve. Maybe there's nothing we can do.

Thank you once again, that’s really very nice.

 

 

Franck VEDEL

 

 

 



Le 3 janv. 2024 à 12:26, Oliver Weinmann <oliver.weinmann@me.com> a écrit :

 

Hi Franck,

 

Great. Glad it works. That is the k8s version and the url where magnum will pull the container from. The official docs are quite helpful. In the latest Openstack release bobcat there is already some support for CAPI. It is not 100% working out of the Box but it looks very promising. I have taken my blog post offline since I didn’t manage to get it working. But I will look into it again in the next couple of days.

 

Cheers,

Oliver

 

Von meinem iPhone gesendet



Am 03.01.2024 um 12:21 schrieb Franck VEDEL <franck.vedel@univ-grenoble-alpes.fr>:

 Hi Oliver,

and thanks a lot for your help.

I followed the instructions on the blog and indeed, with [trust] it works.

I'm trying to understand everything. what are the parameters for:

 

--labels kube_tag=v1.21.11-rancher1,hyperkube_prefix=docker.io/rancher/

 

Thanks in advance

 

Franck

 

 



Le 25 déc. 2023 à 19:26, Franck VEDEL <franck.vedel@univ-grenoble-alpes.fr> a écrit :

 

Hi Oliver and  thanks you for this help.

 

I was wrong in my explications.

I had seen this article (roksblog.de ) and I did what it said:

I first tried:

enable_cluster_user_trust = True  in globals.yml

It didn't work.

 

I then made the modification in /etc/kolla/config/magnum.conf:

cluster_user_trust = True

 

Same problem.

 

That's why I asked the question. Is this really the way to do it? 

If so, I will have to restart the containers (keystone?)



Franck

 

 



Le 25 déc. 2023 à 18:40, Oliver Weinmann <oliver.weinmann@me.com> a écrit :

 

Hi Frank,

 

According to the official kolla docs the option needs to be in globals.yml

 

Ensure that you understand the consequences before enabling this option. Inglobals.yml:

 

In case this also doesn’t work, you can try:

 

yoga) [vagrant@seed ~]$ cat /etc/kolla/config/magnum.conf

[trust]

cluster_user_trust = True

 

*www.roksblog.de

 

Cheers,

Oliver

 

Von meinem iPhone gesendet



Am 25.12.2023 um 18:10 schrieb Franck VEDEL <franck.vedel@univ-grenoble-alpes.fr>:

enable_cluster_user_trust = True