Failed to run a custom cluster-api-provider-baremetal about metal3-dev-env HOT 5 CLOSED

Thanks for the report.

This is actually the expected behavior. make run runs the component in the foreground so that you can follow the log output in the terminal and quickly kill and restart as needed. You will need to interact with the cluster from another terminal.

Let me know if you have any other questions about this!

from metal3-dev-env.

Thanks @russellb.

Indeed you are right. Since, I am new to this environment, I still have one more question.

As mentioned, in my report, I followed the steps mentioned in README. At first, I scaled down the custom Machine controller manager first using the following command

kubectl scale statefulset cluster-api-provider-baremetal-controller-manager -n metal3 --replicas=0.

I added few lines of log prints in my local cluster-api-provider-baremetal-controller-manager repository(in actuator .go) and did the make run. And then when I made changes in the cluster (I was trying to delete a machine object), I saw the new log prints in the console.

However, what I expected is that make run will scale up with the newly compiled cluster-api-provider-baremetal-controller-manager and that shall be reflected when I list the pods. Which is not the case here. So, where is the newly compiled cluster-api-provider-baremetal-controller-manager running since I cannot see this when I list the pods in the cluster ?

Thanks in advance.
Kashif

Thanks for the report.

This is actually the expected behavior. make run runs the component in the foreground so that you can follow the log output in the terminal and quickly kill and restart as needed. You will need to interact with the cluster from another terminal.

Let me know if you have any other questions about this!

from metal3-dev-env.

When you use make run, you are no longer running the component in the cluster, so you won't see it when you list pods. To run a custom image that way, you'd have to build a custom image, push it to a container registry, edit the image in the appropriate manifest under the config/ directory, and then run make deploy. It's a lot easier to test custom changes with make run, just at the expense of it no longer running in the cluster.

from metal3-dev-env.

(sorry, didn't mean to click close)

from metal3-dev-env.

Thanks @russellb. This is clear now.

from metal3-dev-env.