Hello,
I have my first test now with Portworx in Openshift Origin (OKD 3.11)
I created the specs to use Portworx etcd in US, testing is being done in AWS, us-east North Virginia region.
According to screenshots attached, it is possible to notice the folowing items are ready and running :stork, api , px-pvc controller) but px nodes are not ready.
What could be the reason ? Master node and the 3 work nodes are all ready
Best regards,
Paulo
Hello Paulo,
Did you provide disks to Portworx to use?
Could please send us logs from one of the Portworx pods that are not ready?
Regards,
Alex.
Hi Alex,
Yes, I provided /dev/nvme2 disks as you can see in the log attached
In fact, log also shows Portworx stated “he could not find the node in the Kubernetes cluster” (see attached screenshot). Have you seen this before?
Best regards
What is the output of following commands?
kubectl get nodes -o wide
kubectl describe svc kubernetes
It seems your nodes cannot reach the master.
How did you create the kubernetes cluster in AWS?
Regards,
Alex.
Hello Alex,
See commands output attached.
I created this okd cluster based on a “cookbook” from a Dutch consulting company Rubix. It is an environment exclusive for testing purposes. Because it is very difficult to set up okd 3.11 and all that ansible stuff in AWS, they have an approach to make things easier for testing environments, creating one node “all in one” and then scale-up the cluster using scaleup.yaml to the number of compute nodes you want, after you disable scheduling on master node so he becomes infra_master. As I can only put one web link or attachment per forum message, I cannot send directly rubix links.
P.S.: Are you Alex Diaz from Mexico? If yes, maybe you remember me, Paulo from Alcore Brazil, now I am working as an independent consultant
Best regards,
Hi Paulo,
No I am not Alex Diaz. 
I am the other Alex in Portworx (Alex Da Silva - originally from Brazil).
The problem is likely kubernetes network configuration or security group.
Could you please check your security group and see if worker nodes can reach the master?
Per output below kubernetes api is listening on port 8443 and DNS is configured on port 8053.
Regards,
Alex
Hi Alex,
Many thanks, seems it was this settings - my mistake, there was only 53 with traffic allowed in security group!
So you are Brazilian like I am. From Sao Paulo?
Thank you (Obrigado) See you later (Até logo)!
Best Regards (Saudações)
Paulo
Hi Paulo,
I am glad it is working now. My pleasure to help.
I grew in Sao Paulo state, but have been leaving in US for some time now.
Let us know if you need anything else.
Tenha um bom fim de semana.
Alex.
Alex,
Just to update, due to okd/origin problem where it is requested
from Portworx to disable Network Manager, but OKD 3.11 needs
Network Manager, it is part of the pre-install check I installed kops on AWS and everything works
fine.
I just need to test new PX-Backup and see how it
works.
Best regards,
Paulo