-
Notifications
You must be signed in to change notification settings - Fork 1.4k
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
Replace references to tectonic with openshift #644
Replace references to tectonic with openshift #644
Conversation
@wking had a previous (can't remember where) suggestion to keep stuff void of both tectonic / openshift esp in tf templates... |
I don't remember where either ;). But yeah, I think it's best to leave what-we're-installing naming out internally (where there should be no parallel installers for other things that we need to distinguish ourselves from). We just need |
data/data/aws/master/main.tf
Outdated
@@ -96,7 +96,7 @@ resource "aws_instance" "master" { | |||
tags = "${merge(map( | |||
"Name", "${var.cluster_name}-master-${count.index}", | |||
"kubernetes.io/cluster/${var.cluster_name}", "owned", | |||
"tectonicClusterID", "${var.cluster_id}", | |||
"openshiftClusterID", "${var.cluster_id}", |
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.
whoa, boy ;). This one is quasi-public, so I wouldn't be surprised if we break some third-party tooling with this change. Still, I think dropping tectonic
makes sense to get consistent branding. And openshift
is probably appropriate here, although I'd be open to "kubernetes.io/cluster-id", ${var.cluster_id}
if we could drum up an appetite for a cluster UUID tag at the vanilla Kubernetes level.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have removed the change of tectonicClusterID
from this PR. I will do it in a subsequent PR.
data/data/aws/route53/openshift.tf
Outdated
@@ -3,17 +3,17 @@ locals { | |||
private_endpoints_count = "${var.private_endpoints ? 1 : 0}" | |||
} | |||
|
|||
data "aws_route53_zone" "tectonic" { | |||
data "aws_route53_zone" "openshift" { |
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.
Potential generic name here would be base
to match base_domain
. I'd also be fine with zone
or cluster
or similar. I'm not terribly opposed to openshift
here though, if folks prefer that to my alternatives.
|
||
[Service] | ||
WorkingDirectory=/opt/openshift/openshift | ||
ExecStart=/usr/local/bin/openshift.sh /opt/openshift/auth/kubeconfig |
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.
I'd prefer a more specific name for this script and service, since "OpenShift" feels like "Kubernetes plus the OpenShift-specific stuff" and not "just the OpenShift-specific stuff" to me. The script is about pushing our manifests into the cluster, so maybe manifests.sh
(and manifests.service
)? Or initial-manifests.sh
? operator-manifests.sh
? Or something? special-sauce-beyond-vanilla-kubernetes.sh
? ;)
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.
I had changed it to manifests.service
and put the manifests that it used at /opt/openshift/manifests
, but there were conflicts with manifests coming from another location. I was not in love with the name manifests.service
as there are a lot of different manifests used for different purposes. If we really care about changing this name, then we can have a discussion around better names and change it in another PR.
data/data/aws/main.tf
Outdated
public_endpoints = "${var.tectonic_aws_endpoints == "private" ? false : true}" | ||
private_zone_id = "${var.tectonic_aws_external_private_zone != "" ? var.tectonic_aws_external_private_zone : join("", aws_route53_zone.tectonic_int.*.zone_id)}" | ||
private_endpoints = "${var.openshift_aws_endpoints == "public" ? false : true}" | ||
public_endpoints = "${var.openshift_aws_endpoints == "private" ? false : true}" |
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.
Besides tectonic
/openshift
being redundant. aws
is also redundant (it's implied by the module name). I'd be fine with just ingress
or some such.
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.
I agree that aws
is redundant. However, the aws
is needed by the json marshal/unmarshal to distinguish between fields in AWS
, Libvirt
, and OpenStack
in the pkg/tfvars/config
struct in case there are fields that would otherwise have identical names. This is a side effect of inlining all of those fields in pkg/tfvars/config
.
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 is a side effect of inlining all of those fields in
pkg/tfvars/config
.
I think Terraform needs a single level. But we only ever need one platform at a time. We could (future work?) invert the structs to embed the generic struct within each per-platform struct.
This is also pretty big. It might be easier if we try to bite it off in smaller chunks? E.g. just stuff inside Terraform in one PR. Then the Go -> Terraform JSON serialization in a second PR. Then the pure Go stuff in a third PR? |
i would say just separate commits is fine.. separate PR is too much overhead. |
I just launched a libvrit cluster from febbd75, and it came up without issues. And I'm super excited about getting this landed, just needs a rebase now (and one typo fix, although I'm fine landing this without that fix too). |
The tectonic prefix in the tfvars are redundant as it identifies the type of installer being used which is evident already by using that installer. Changes for https://jira.coreos.com/browse/CORS-878
This makes the corresponding change in the terraform to match the removal of the terraform prefix in the tfvars json. Changes for https://jira.coreos.com/browse/CORS-878
The session name when connecting to AWS is changed from starting with TECTONIC_INSTALLER_ to starting with OPENSHIFT_INSTALLER. The terraform name of the internal route53 zone is changed from terraform_int to int. The Name tag of the zone is changed from ending with _tectonic_int to ending with _int. The terraform name of the base route53 zone is changed from tectonic to base. The terraform name of the api route53 records are changed from tectonic_api_external and tectonic_api_internal to api_external and api_internal, respectively. Changes for https://jira.coreos.com/browse/CORS-878
The terraform name of the libvirt network is changed from tectonic_net to net. Changes for https://jira.coreos.com/browse/CORS-878
…ules The terraform name of the openstack object storage container is changed from terraform to container. Changes for https://jira.coreos.com/browse/CORS-878
Change the directory of bootstrap files from /opt/tectonic to /opt/openshift. Change the directory of files laid down by the Tectonic Manifests asset from tectonic to openshift. Change the name of the service that lays down OpenShift manifests from tectonic.service to openshift.service. Changes for https://jira.coreos.com/browse/CORS-878
@staebler: The following test failed, say
Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
Change directory for openshift manifests from data/data/manifests/tectonic to data/data/manifests/openshift. Chnages for https://jira.coreos.com/browse/CORS-878
…tonic to manifests/openshift Changes for https://jira.coreos.com/browse/CORS-878
Bazel is no longer used. Changes for https://jira.coreos.com/browse/CORS-878
/lgtm |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: staebler, wking The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Looks like the job might be hung in tests? @staebler, can you grant me access to ci-op-rk8i4bxq or poke around there yourself and see if anything looks fishy? |
Changes for https://jira.coreos.com/browse/CORS-878.
There are still two types of references to tectonic.
tectonicClusterID
. I will change this toopenshiftClusterID
in a subsequent PR. The change requires corresponding changes in other repos. I wanted to isolate the change so that the changes in this PR are not held up.