Search results

Thank's, Me too. I spoke about this problem on Twitter last night with Alexandre. Moula.

I migrated one of my datacenters to PVE 8.0 beta with Ceph, everything went well!!! Even my Openstack, Kubernetes, Kubeflow, Mlflow, maas... bundles work perfectly!!! good job!!!

MAAS can do it. Just add your pve iso with Packer as it is done with vmware, rocky-linux... https://github.com/canonical/packer-maas

A single Static public IP I recommend you tu use a "Reverse Proxy" https://nginxproxymanager.com/ https://doc.traefik.io/traefik/getting-started/quick-start/ etc...

One of my data centers is based on this technology (image attached ) with a network dedicated directly to ceph. But above I go through a VM in H.A as a reverse-proxy for the projects: nginx-reverse-proxy ( Letsencrypt) +freeipa+authelia for security. https://nginxproxymanager.com/...

Suggestion : Why not use the module SDN : libpve-network-perl ? With the latest version @spirit "Alexandre" has done a good job. Here is a link if you want to follow it. https://blog.raspot.in/fr/blog/mise-en-place-du-sdn-sur-promox-7 Sorry it's in french. Moula.

I tested the SDN module and it works fine. I have two comments. 1- The menu display (see p1). 2- I want to delete everything but no possibility I have the message (see p2) while I have deleted all the vms that I had deployed with !!! I will continue to test its integration with Netbox and...

It is all good. I updated each node individually, without Cluster, everything worked. The slowness came from updating the cluster with the pve-cluster package. Thank's.

Bonjour, With the new packages of this morning, the update is not done. It does not go to the end on all the nodes. I'll see where the problem comes from.

Bonjour, I've been making the point to the Canonical team for a while. But still no solution, we must already modify Centos with Rocky-Linux (Alma-Linux). I have seen recently with the arrival of a new developer (Christian Grabowski) a lot of new things in this direction. I think fixing the...

Bonjour, Sorry, I don't see any shared storage !!! To do H.A you need a shared storage like Ceph on each node of your data center or another shared storage like PBS, TRUENAS ... with the ISO or CT on it.

No, it's still not possible to create proxmox nodes. This is a compatibility issue of the Debian network with the Curtin networks of ubuntu. But MAAS allows to create, manage, deploy vms on Proxmox and authentication with Tokens very well.

Bonjour, With MAAS + JUJU no need for ansible + Terraform. Everything is automated with the bundles and image-charmed ... Whether on Proxmoxve with Vms or on baremetal.. https://charmhub.io/ https://jaas.ai/search?q=openstack&type=bundle ( Bundle openstack )...

Hi, Sorry, I don't know anything about VMWARE. I can just tell you the goal : you can create as many clouds as you want for each "user group" for cloud-computing, fog computing, Edge Computing, Hpc ... with the K8s, openstack, microk8s , microstack, hkupe, Kubeflow ... and all on a PROXMOXVE +...

Hi, I see that your GPU card is detected in PCI-E. and well configured. Have you tried installing a Gnu/linux distribution ( Debian or Ubuntu for example) with the Nvidia drivers? I know absolutely nothing about windows :D Remember that you can only use one vm with GPU. Thank's.

I think the problem is with your GPU, I can't see anywhere where you talk about GPU added to your motherboard. The problem comes from your GPU soldered to your motherboard. Add an Nvidia GPU to your servers and you'll see that everything will work fine. Look at my Datacenter with Nvidia GPUs...

Deploy your services on the cloud edge in a few clicks, all automated with open-source projects: Proxmoxve+maas+juju+microk8s+portainer...

I create as many Openstacks as I want with the bundles. Here is a link of my lessons (Practices) with my Master2 students ... https://github.com/DOSSANTOSDaniel/MAAS_JUJU_K8S_Openstack If that can help you. Since your questions concern more openstack than ProxmoxVE, please contact me outside...

This "Network" part I manage it before the creation of juju-controller with the MAAS project. even OVS is integrated in the dev version.