Changes to deploy baremetal-operator within the cluster created by dev-scripts. #139
Conversation
08_deploy_bmo.sh
Outdated
| oc --as system:admin --config ocp/auth/kubeconfig apply -f deploy/crds/metalkube_v1alpha1_baremetalhost_crd.yaml | ||
|
|
||
| export OPERATOR_NAME=baremetal-operator | ||
| operator-sdk up local --namespace=bmo-project |
There was a problem hiding this comment.
This is going to run the operator on the local host, and block the installer. You can replace lines 15 and 16 with something like
oc --as system:admin --config ocp/auth/kubeconfig apply -f deploy/operator.yaml
That will launch the most recent version of the operator using the image referenced in the operator.yaml file.
|
Build SUCCESS, see build http://10.8.144.11:8080/job/dev-tools/154/ |
|
I think you can remove the WIP from the commit message. You mentioned wanting to update the README, and I think that's good. As we discussed, you should update the Makefile, too, so that the operator is always installed. |
|
Build SUCCESS, see build http://10.8.144.11:8080/job/dev-tools/160/ |
bmo_install_prereqs.sh
Outdated
| @@ -0,0 +1,13 @@ | |||
| # Install operator-sdk for use by the baremetal-operator | |||
There was a problem hiding this comment.
This could probably go in 03_ocp_repo_sync.sh where we clone various other things, fine to leave it as a separate file if you prefer though
08_deploy_bmo.sh
Outdated
| source bmo_install_prereqs.sh | ||
| # Get the latest bits for baremetal-operator | ||
| go get github.com/metalkube/baremetal-operator | ||
| export BMOPATH=~/go/src/github.com/metalkube/baremetal-operator |
There was a problem hiding this comment.
And similar to the comment below we could use 03_ocp_repo_sync.sh to get the repo vs go get, the original idea was that we'd clone everything needed there so that test iterations on the other scripts during development could be faster.
|
Incorporated all review comments. Running commands within 08_deploy_bmo.sh are resulting in the following error(s): After tests run successfully:
|
|
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/177/ |
sadasu
changed the title
|
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/178/ |
|
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/179/ |
|
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/180/ |
|
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/187/ |
|
Build SUCCESS, see build http://10.8.144.11:8080/job/dev-tools/190/ |
hardys
left a comment
hardys
left a comment
There was a problem hiding this comment.
I tested this with the in-progress fixes for the cluster bootstrapping, and it works other than the missing GOPATH mentioned in the comments:
[shardy@tripleodev2 dev-scripts]$ ./08_deploy_bmo.sh
Error from server (AlreadyExists): project.project.openshift.io "bmo-project" already exists
serviceaccount/baremetal-operator created
role.rbac.authorization.k8s.io/baremetal-operator created
rolebinding.rbac.authorization.k8s.io/baremetal-operator created
customresourcedefinition.apiextensions.k8s.io/baremetalhosts.metalkube.org created
deployment.apps/baremetal-operator created
If you can fix that, I think we can merge this and soon the workarounds for cluster bootstrapping should land to enable folks to test with it.
|
Other than to GOPATH issue I mentioned I think this is good to go, @dhellmann any concerns with merging after that gets fixed? |
|
I'm happy to go ahead with this when the GOPATH thing is fixed. |
created by dev-scripts.
|
Both GOPATH and the GOPATH/bin were added in my profile and so didn't run into these issues locally. Thanks for catching it. |
Ok thanks, FYI now that the #159 workaround landed it should be possible to test this when the GOPATH issue is fixed |
|
Build FAILURE, see build http://10.8.144.11:8080/job/dev-tools/209/ |
4 participants
No description provided.