Page MenuHomeSoftware Heritage

Upgrade vm templates
Closed, MigratedEdits Locked

Description

When provisioning new rancher cluster or new nodes, vm templates are old enough that it takes at least one hour [1] to provision one node.
It's too long.

Updating template should cut down that time by some unknown order of magnitude.

To reproduce, just clone out of the proxmox ui our template 10008 (bullseye-zfs) and start it.
The boot takes forever as cloud-init is upgrading the zfs part (kernel module compilation and whatnot).
That must be the pain point which should disappear altogether if we upgrate that template.

[1] D8155#212105

Event Timeline

ardumont triaged this task as Normal priority.Jul 27 2022, 12:42 PM
ardumont created this task.
ardumont changed the task status from Open to Work in Progress.Jul 27 2022, 12:45 PM
ardumont claimed this task.
ardumont updated the task description. (Show Details)
ardumont moved this task from Backlog to in-progress on the System administration board.

Vagrant templated got updated and Vagrantfile declaration as well, see commit ^.

For proxmox:

  • bullseye template [1]
  • bullseye with zfs template [2]

Ongoing commits with the proper impact on the terraform manifests.

[1] P1412

[2] P1413

After creating a new graphql-worker3, it took only 36min [1] instead of 1h15min... So, a
tad faster.

It's still slow though...

[1]

module.graphql-worker3.proxmox_vm_qemu.node: Creation complete after 36m2s [id=uffizi/qemu/165]

Apply complete! Resources: 1 added, 0 changed, 0 destroyed.