Provisioning: Add Oracle Cloud Infrastructure #652
Conversation
|
Draft as it's not complete yet. I still need to figure out the networking section. |
| ---- | ||
| source_image"fedora-coreos-qemu-image-name.qcow2" | ||
| image_name="fcos-oraclecloud.qcow2" | ||
| ./covert-image.sh "${source_image}" "${image_name} oraclecloud |
There was a problem hiding this comment.
I used openstack as a workaround until we get support in order in Ignition.
This uses the legacy IMDS endpoint (v1) so it must not be disabled when creating the image.
|
I'm sure you've seen this, though note that bare metal and VM shapes have different requirements: coreos/fedora-coreos-tracker#414 (comment). Short-term, getting the VM shapes to work shouldn't be too much work. For the bare metal shapes, we could for now recommend using the bare metal installation flows, until we have "dual" support in the same image. |
|
Ah indeed, thanks for the pointer. So those docs would be only for VM shapes right now. I should update the list of shapes to select only those. For Bare Metal shapes, maybe we should add the kargs directly to the image as part of the conversion step that we have right now as we don't produce oracle cloud specific images. Then we would have users upload this image and select the right shape for it. |
travier
force-pushed
the
main-prov-oraclecloud
branch
2 times, most recently
from
00e785b to
e2daba8
Compare
I think that's fine for now, yeah. |
travier
force-pushed
the
main-prov-oraclecloud
branch
3 times, most recently
from
f6c9b86 to
7eef04d
Compare
|
Based on top of #654 for convenience as its likely to be merged earlier. |
travier
force-pushed
the
main-prov-oraclecloud
branch
from
7eef04d to
cacaed1
Compare
| .QEMU to Oracle Cloud Infrastructure image conversion script | ||
| [source, bash] | ||
| ---- | ||
| #!/bin/bash |
There was a problem hiding this comment.
Definitely would be nice to just stick this in a container image, but there's overhead there too in maintenance.
AsciiDoc supports including files in multiple documents. That would allow us to define this script just once to make it easier to maintain here. (And I guess make the plaform ID be another parameter.)
There was a problem hiding this comment.
Agree that it would be nice to find a good place for this script.
travier
force-pushed
the
main-prov-oraclecloud
branch
3 times, most recently
from
3fa6537 to
91d55f5
Compare
|
i tried your instructions, but it seems that the metadata endpoint used in afterburn is not available here: Jul 19 12:42:30.959 WARN failed to locate config-drive, using the metadata service API instead Instead i found it available here: http://169.254.169.254/openstack/latest/meta_data.json or here http://169.254.169.254/opc/v1/instance/metadata/ssh_authorized_keys ist there a way to overide the metadata endpoint that is used by afterburn (i.e by kernel command line)? |
|
Thanks for giving it a try. This is due to the fact that we re-use the openstack implementation as is for now which is not the long term plan: coreos/fedora-coreos-tracker#414 (comment) This is why I've commented out all Afterburn mentions in the docs for now. As far as I know, afterburn does not let you configure that on the kernel command line. |
Initial documentation to setup FCOS on Hetzner. Inspired by: https://www.flatcar.org/docs/latest/installing/cloud/hetzner/ See: coreos/fedora-coreos-tracker#1324
Initial documentation to setup FCOS on Oracle Cloud Infrastructure. See:coreos/fedora-coreos-tracker#414
travier
force-pushed
the
main-prov-oraclecloud
branch
from
91d55f5 to
f84d417
Compare
Initial documentation to setup FCOS on Oracle Cloud Infrastructure.
See:coreos/fedora-coreos-tracker#414