Machine not finishing deployment | Node installation failure - 'cloudinit' running modules for final

Users
,
#1

Hey everyone,
I keep receiving this error even though I am successfully SSHed into the MAAS-provisioned machine. Any idea why this is happening and how I can solve it?

I have seen this bug report but I am not sure how to apply it and if it relevant for my arm64 case. Do I just copy and paste this code into the cloud-init script during re-deployment?

I have attached some logs from the machines as well. Thanks for your time and attention.

Region+Rack Controller
cat /var/snap/maas/common/log/regiond.log
cat /var/snap/maas/common/log/maas.log

Rack Controller
cat /var/snap/maas/common/log/rsyslog.log
cat /var/snap/maas/common/log/rackd.log
cat /var/snap/maas/common/log/maas.log

calm-fox machine error logs from MaaS UI

image
image857×570 17.4 KB

Machine not finishing deployment and changing status after deployment.

image
image1131×514 29.5 KB

#2

My experience for arm 64 (Raspberry Pi) is that only focal with hwe min kernel works. I get various similar errors to yours for other OS versions.

#3

I am currently using Ubuntu 20.04 LTS focal here on all my arm64 PIs that’s why I am confused why I’m receiving this error.

#4

So is this happening on one machine only? You can get more detailed logs with the Download button on the logging pane, does that tell you anything more?

#5

Yes only on one machine.

image
image1050×103 5.85 KB

This is the only error I see here since machine finishing Configuring OS.

I have attached logs below as well.
calm-fox.maas-machine-output-2022-03-10.xml
calm-fox.maas-machine-output-2022-03-10.yaml
calm-fox.maas-installation-output-2022-03-10.log
Somehow it seems MaaS can’t communicate after deployment to clear the misunderstanding up, because the node is working fine connected as a microk8s master node to the rest of my arm64 cluster.

EDIT: Also I noted that KVM Host is not detected for this machine.

image
image728×441 14.1 KB

#6

Sorry, not sure what is happening here. Since the curtin logs look fine.

I’m not experienced with KVM so can’t comment on that.

My only suggestion is to release the machine redo the commissioning, deployment. (but you probably already tried that!)

I don’t think that the fix you linked is likely to help since it is quite old. However if you want to give it a go you need to apply the patch to the file compose_preseed.py. For details on how to do that, see https://www.howtogeek.com/415442/how-to-apply-a-patch-to-a-file-and-create-patches-in-linux/

#7

@ryzengrind, see this other discourse post. does anything from there give you a clue?