Talos on Hyperstack cloud

[kubebn]

I want to try to deploy Talos on top of Hyperstack, but I’m currently struggling to determine the best process to follow.

From what I understand, the image can be created either by taking a snapshot of the current VM or by using a bootable volume.

I was initially considering using the existing documentation for Hetzner as a reference — for example: https://www.talos.dev/v1.9/talos-guides/install/cloud-platforms/hetzner/. However, it seems that this approach might not be applicable to Hyperstack, since it lacks support for rescue mode, Packer, or similar utilities.

At this point, it looks like the only potentially viable approach is the one I shared here: #10737. That said, I’m still unclear on how to proceed in a way that allows me to apply machine configurations via user-data, or a similar mechanism.

Any advice or suggestions would be greatly appreciated. If this approach ends up working, I’d be more than happy to contribute a guide for the Talos documentation.

[rothgar]

I don't know anything about Hyperstack but it doesn't appear you can upload your own disk image to their platform. In that case you can try the steps to "take over" an ubuntu system with the steps here

#10737 (comment)

These are untested by us and you probably don't need all of the static network config but if you're able to kexec to talos, install talos, and then reset the config (I forget the flag) you should be able to snapshot the volume and re-use that for other VMs.

[kubebn]

I don't know anything about Hyperstack but it doesn't appear you can upload your own disk image to their platform. In that case you can try the steps to "take over" an ubuntu system with the steps here

#10737 (comment)

These are untested by us and you probably don't need all of the static network config but if you're able to kexec to talos, install talos, and then reset the config (I forget the flag) you should be able to snapshot the volume and re-use that for other VMs.

kexec didn’t work; however, using GRUB and the bootloader menu was actually quite helpful. I was able to overwrite Ubuntu and successfully boot into Talos.

I do have one question: whenever I run a reset on the Talos machine using the following command:

talosctl reset -n 10.0.0.156 --reboot --graceful=false --system-labels-to-wipe STATE --system-labels-to-wipe EPHEMERAL

(where the initial configuration was applied via talosctl apply-config), the machine reverts to the state where it’s waiting for the apply-config command again.

I took a snapshot of the machine in this state, prepared an image, and tried creating a VM from it. I also provided user-data with the Talos machine configuration, but the config is not being applied—it just boots and waits for apply-config again.

Is there any way to use cloud-init/user-data in this scenario?

Am I missing something? Is there a limitation that prevents user-data from working with bare metal or custom VM setups?

[smira]

Am I missing something? Is there a limitation that prevents user-data from working with bare metal or custom VM setups?

What is the user-data that you applied? If you're using nocloud platform (for example), you should use nocloud Talos image.

[kubebn]

I made it via:

#cloud-config

write_files:
  - path: /etc/grub.d/10_talos
    permissions: '0755'
    content: |
      #!/bin/sh
      exec tail -n +3 $0
      
      menuentry "Talos" {
          linux /boot/talos-kernel init_on_alloc=1 slab_nomerge pti=on console=tty1 console=ttyS0 consoleblank=0 nvme_core.io_timeout=4294967295 printk.devkmsg=on ima_template=ima-ng ima_appraise=fix ima_hash=sha512 talos.platform=openstack net.ifnames=0 ip=${IP}::${GATEWAY}:${NETMASK}::eth0:off talos.dashboard.disabled=1
          initrd /boot/talos-initrd.xz
      }
      menuentry "Reset Talos" {
          linux /boot/talos-kernel init_on_alloc=1 slab_nomerge pti=on console=tty1 console=ttyS0 consoleblank=0 nvme_core.io_timeout=4294967295 printk.devkmsg=on ima_template=ima-ng ima_appraise=fix ima_hash=sha512 talos.platform=openstack net.ifnames=0 ip=${IP}::${GATEWAY}:${NETMASK}::eth0:off talos.dashboard.disabled=1 talos.experimental.wipe=system:EPHEMERAL,STATE
          initrd /boot/talos-initrd.xz
      }

runcmd:
  - sudo -i
  - IP=$(ip -o -4 route get 8.8.8.8 | awk -F"src " '{sub(" .*", "", $2); print $2}')
  - GATEWAY=$(ip -o -4 route get 8.8.8.8 | awk -F"via " '{sub(" .*", "", $2); print $2}')
  - ETH=$(ip -o -4 route get 8.8.8.8 | awk -F"dev " '{sub(" .*", "", $2); print $2}')
  - CIDR=$(ip -o -4 addr show "$ETH" | awk -F"inet $IP/" '{sub(" .*", "", $2); print $2; exit}')
  - NETMASK=$(echo "$CIDR" | awk '{p=$1;for(i=1;i<=4;i++){if(p>=8){o=255;p-=8}else{o=256-2^(8-p);p=0}printf(i<4?o".":o"\n")}}')
  - sed -i "s/\${IP}/$IP/g; s/\${GATEWAY}/$GATEWAY/g; s/\${NETMASK}/$NETMASK/g" /etc/grub.d/10_talos
  - sed -i 's/^GRUB_DEFAULT=.*/GRUB_DEFAULT=2/' /etc/default/grub
  - wget -O /boot/talos-kernel https://github.com/siderolabs/talos/releases/download/v1.9.5/vmlinuz-amd64
  - wget -O /boot/talos-initrd.xz https://github.com/siderolabs/talos/releases/download/v1.9.5/initramfs-amd64.xz
  - update-grub
  - sleep 2
  - reboot

[smira]

Talos doesn't support generic cloud-init, as it doesn't have any kind of shell to execute the commands (and it doesn't have most of the commands as well).

When running on bare-metal, you can e.g. use talos.config= to supply Talos machine configuration.

You can use Omni to manage machine configuration as well.

[kubebn]

Talos doesn't support generic cloud-init, as it doesn't have any kind of shell to execute the commands (and it doesn't have most of the commands as well).

When running on bare-metal, you can e.g. use talos.config= to supply Talos machine configuration.

You can use Omni to manage machine configuration as well.

Yeah, for metal I was usually using this https://www.talos.dev/v1.9/reference/kernel/#metal-iso