Some VMs Boot into initramfs

[3deep5me]

/kind bug

What steps did you take and what happened: I create a cluster as described in the quick-start. I changed the replicas to 3 for md and controlplane. I also changed the storage for the vm-disks to a lvm-thin storage. Some VMs (in this example two and three) do not boot correctly instead there a going into the initramfs. Therefore the machines does not go into running state. Its random which VMs run into the error. Sometimes its even the first controlplane VM.

first try

k get machines
NAME                           CLUSTER       NODENAME                         PROVIDERID                                       PHASE         AGE     VERSION
cappx-test-6dqst               cappx-test    cappx-test-controlplane-llwfx    proxmox://c6244ce6-be79-4073-a378-8b0fd75a79cf   Running       6m47s   v1.27.3
cappx-test-kw42b               cappx-test    cappx-test-controlplane-zj9b7    proxmox://81616cfd-df41-4d1a-aeab-c0fdb76c2987   Running       10m     v1.27.3
cappx-test-md-0-cnslp-45tvt    cappx-test                                     proxmox://a15e2fb3-5d6f-4b1f-84fd-7fb49c91abf8   Provisioned   13m     v1.27.3
cappx-test-md-0-cnslp-c28nt    cappx-test                                     proxmox://1c38b356-fa5a-4343-8cd9-e28f013cd55f   Provisioned   13m     v1.27.3
cappx-test-md-0-cnslp-td5hh    cappx-test    cappx-test-md-0-2fmq7            proxmox://6cf85bbb-c1d8-46df-a38e-b5b6f25dbf68   Running       13m     v1.27.3
cappx-test-tnrmr               cappx-test    cappx-test-controlplane-4fds9    proxmox://dc4da772-23a6-4446-80e8-40527dc8ac55   Running    

second try

k get machines
NAME                           CLUSTER       NODENAME                         PROVIDERID                                       PHASE         AGE     VERSION
cappx-test-cbbp8               cappx-test    cappx-test-controlplane-gkpbx    proxmox://cea53be2-dd58-4902-80c5-3eab7ba9e3d5   Running       3m17s   v1.27.3
cappx-test-md-0-85954-9ctbs    cappx-test                                     proxmox://3c92d086-7bce-4f4e-8986-2e1d43629aa5   Provisioned   6m51s   v1.27.3
cappx-test-md-0-85954-g6lnk    cappx-test                                     proxmox://0d6f79a8-867d-4a79-af75-3b7970dc498d   Provisioned   6m51s   v1.27.3
cappx-test-md-0-85954-j82jv    cappx-test                                     proxmox://5764bd21-660c-42aa-9466-bf4fe321cbeb   Provisioned   6m51s   v1.27.3
cappx-test-qdpkj               cappx-test    cappx-test-controlplane-v92sr    proxmox://3e631789-3aa5-4e05-846f-31e3a6a0a857   Running       6m50s   v1.27.3

What did you expect to happen: The VMs should start as the others.

Anything else you would like to add: I'm not sure if this is cappx related if not maybe this is something which can maybe be caught by a MachineHealthCheck. (Will investigate in this)

Environment:

• Cluster-api-provider-proxmox version: 0.3.3
• Proxmox VE version: Virtual Environment 8.0.4
• Kubernetes version: (use kubectl version):

Client Version: v1.28.2
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
Server Version: v1.27.3

• OS (e.g. from /etc/os-release): debian(client) ubuntu(default, server)