-
Notifications
You must be signed in to change notification settings - Fork 191
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
WIP run Ironic on the bootstrap VM #642
Conversation
TODO is updating the readme/docs to show how to access the baremetal IP of the bootstrap VM with openstackclient. |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
51e22cf
to
e4bb177
Compare
This comment has been minimized.
This comment has been minimized.
374e934
to
1c5fddc
Compare
This comment has been minimized.
This comment has been minimized.
-v $IRONIC_DATA_DIR:/shared --entrypoint /bin/rundnsmasq ${IRONIC_IMAGE} | ||
|
||
# We start only the httpd and *downloader containers so that we can provide | ||
# cached images to the bootstrap VM | ||
sudo podman run -d --net host --privileged --name httpd --pod ironic-pod \ | ||
-v $IRONIC_DATA_DIR:/shared --entrypoint /bin/runhttpd ${IRONIC_IMAGE} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This appears to be getting used during the deploy (IPA is downlownloading it from this container), I guess the URLs need to change here
dev-scripts/make-bm-worker/main.go
Line 13 in d540fd8
instanceImageSource = "http://172.22.0.1/images/rhcos-ootpa-latest.qcow2" |
and
dev-scripts/ocp_install_env.sh
Line 80 in d540fd8
source: "http://172.22.0.1/images/$RHCOS_IMAGE_FILENAME_LATEST" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ah yeah good point, I'll change it in the install-config for now - ultimately I'd like to remove the image from there and just reference it inside the installer, fixing openshift-metal3/kni-installer#58
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok so it turns out this won't work, because the image URL/checksum passed via the install-config is used for both terraform to deploy the masters, and also to set the Image on the providerSpec for the worker machineset.
So I think we'll have to fix openshift-metal3/kni-installer#58 and remove that install-config interface, such that we can internally decouple the two different image locations. Digging into that now.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
latest version in openshift/installer#2079 should resolve this I think
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This doesn't seem to work, is the idea the bootstrap should get the image from the httpd pod running on the provisioning host? It's currently downloading from the openshift URL, which is very slow from some of our boxes. In fact, on my hardware in one of data centers, I get a lousy 150kbps, which takes longer than 30 min -- and thus the installer timeout -- to download RHCOS. I'm wondering if that's why we haven't seen this pass in CI.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Local caching wasn't working because the hardcoded RHCOS image in dev-scripts didn't match the new version in the installer. I pushed an update here, so the newer RHCOS does get downloaded and cached.
It will never work in CI as written, because we don't cache the RHCOS images in CI. This is trickier to solve, but I'll try fixing it once I actually get the rest of the stuff working on my local box.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
70bbe2d
to
3b79ca6
Compare
This comment has been minimized.
This comment has been minimized.
3b79ca6
to
c2aeff3
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/995/ |
92a68d8
to
921f4e7
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
921f4e7
to
a514cfc
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Build SUCCESS, see build http://10.8.144.11:8080/job/dev-tools/1003/ |
CI is passing. We'll need a new openshift release once the installer PR is merged. |
🎉 Now we need a way for the CI to retrieve the ramdisk logs (that are stored as tarballs in the shared location). |
This is to test with openshift-metal3/kni-installer#100