Creating multiple VMs in a proxmox cluster, but the first one never boots about terraform-provider-proxmox HOT 7 OPEN

Thanks for the report @Neels-v-Wyk! Definitely, there is something interesting going on, I haven't seen such behaviour before. Would you mind provide some extra details?

If you deploy just one VM, does it work?
Is it always failing on server-01? Could you try reverse order of the servers?
Is there any errors in tasks lists on the cluster?

from terraform-provider-proxmox.

If you deploy just one VM, does it work?

Yup! Setting the count to 1 results in one working vm on server-01. Setting it to 2 results in one broken one on server-01 and a working one on server-02. Interestingly enough, setting the servers to be created by trying to reverse the count 3 - count.index yields the same result, I think probably because terraform doesn't creating things in order but in parrallel, so this would make no difference

Is there any errors in tasks lists on the cluster?

No task errors, no errors in the logs either. It seems that the creation process is working, I'm assuming this is something to do with how the machine starts

Explicitly excluding server-01 form being selected for creating VMs and only creating them on 02 and 03, both still get created correctly :(

...

Alright, I tried a few more things and ended up reinstalling proxmox on server-01 and re joining it to the cluster. Everything is working now, consistently. I wish I had more info on why this worked when creating 1 VM but not multiple. I'll close this issue for now and reopen it if it happens again but it seems like I must have messed something up myself

from terraform-provider-proxmox.

Heck, the issue is back. I can't terraform apply -destroy and then terraform apply again without making it happen. Something that does seem to help though is rebooting the first server, then it works for one instance of terraform apply and a subsequent destroy

from terraform-provider-proxmox.

That is no longer the case ^, I am consistently getting the same issue again :(

from terraform-provider-proxmox.

This seems to be a ZFS related issue. I deleted my ZFS pool and recreated another one, and then nothing started coming up. I deleted my ZFS pool again and extended local-lvm to make use of the freed up space, and now everything works.

I also attempted to use a different proxmox node as the entrypoint but that still resulted on the vm on server-01 not having a boot disk

from terraform-provider-proxmox.

Hi @Neels-v-Wyk!

Was you able to sort things out? Do you need any support from the provider's side?
From what I'm reading here, there seems to be an obscure issue with the Proxmox ZFS storage that prevents VM from a normal boot.

Realistically, I don't think provider / terraform can do anything about it, except maybe set parallelism to 1, so only one VM is provisioned at a time?

from terraform-provider-proxmox.

Interesting! I don't know if that will solve the problem but I think it's a step in the right direction. I'll have to try this later, I'll get back to you if there's any new results

from terraform-provider-proxmox.