Hi,
Since it's the third it's happenning to me, I'll request some help from OVH team about a big issue I got.
I have a managed k8s cluster with 1 node used for CI/CD. All is going well with container creation and running. Then suddenly all pods are out and change to "ContainerCreating" :
I check the nodes :
First hint, the "Age" of 18h seems odd, since I created the node some days ago. Then I check some more info on the node :
Every thing seems here also fine but nothing's running on the cluster ...
When I try to gather some info on a "base service" pod :
So what I suspect is that the node was "re-created" at some point but all local storage and mount points were not correctly transferred.
The only way I have on fixing that is to delete the node and creating one new (with all configurations adjustments needed).
Could you help ?
Thanks
All pods killed and to "Container Creating"
Related questions
- Kubernetes loadbalancer IP source
7662
02.08.2019 06:18
- Pull image kubernetes
7357
08.11.2022 11:07
- Kubernetes dashboard
6884
02.12.2019 14:14
- Issues with Wordpress install guide
6753
13.12.2018 20:11
- Managed Kubernetes Platform is now live !
6266
27.02.2019 15:20
- Certificate order pending
5716
11.09.2023 19:32
- Guides pour la mise en place et l'utilisation des volumes
5415
10.04.2019 12:27
- Erreur intempestive
5382
03.04.2019 17:18
- Mémoire pod limitée java
4894
24.04.2019 22:19
Hi,
It looks like you 'manually' renamed your Public cloud instances after node creation.
This is a known problem that causes issues with node interaction with the control plane (node is unable to access needed resources like tokens, etc)
This is a bit akward to detect it on our side because , as you pointed out, the node is in ready state.
But we are working on the issue, and will communicate once it is resolved.
In the meantime, you are already able to name your nodes at creation when using the OVH API (You will soon be able to do it when using the OVH Manager).
We hope this helps, and are sorry for the inconvenience.
Have a nice day
Hi NicoN,
Thanks for the quick answer. I was really not expecting such an explanation. But the solution's simple : I won't rename my nodes ;)
Thanks a lot