im going through this terrible process of generating a butane file, generating an ignition file, deploying using terraform using custom_data.
if I have something wrong, it just times out at 20 mins or so saying it failed.
I then start removing ignition config until it boots, then process of debugging that config and trying again.
I don’t know of anyway to get what failed in ignition when running it in azure.
Is there something that gives me some insight into ignition failures?
Is there a better way to work?
I would recommend starting with a local VM, using QEMU, etc.
There should also be a way to collect the console logs for Azure VMs.
Enabling boot-diagnostics for the VM in Azure really helps. It lets you grab the serial console logs, which should include ignition errors, systemd failures, etc. You can turn it on with az vm boot-diagnostics enable az vm boot-diagnostics | Microsoft Learn.
One enabled, you can grab the logs with az vm boot-diagnostics get-boot-log az vm boot-diagnostics | Microsoft Learn. You can also view the serial console output directly in the Azure Portal view for the VM under Help → Serial console, which can be easier for quickly scanning through early boot issues. That usually gives enough context to figure out what Ignition tripped on.
2 Likes
thanks both for the response. great tip on the Azure boot diagnostics!
will report back