Tag master machines for adoption by machine controller #479
Conversation
|
@abhinavdahiya ptal |
openshift-ci-robot
added
the
size/XS
|
My idea of adoption was that we create |
|
Are there docs for the |
Yes, that's coming :) Just wanted to get this small change in first. The one blocking issue I have with creating the master Machine assets on install is that I don't have a good way to assign AZ's to each. My understanding is that Terraform resolves those once it is running. However, we will need to know AZ's anyway in order to support Multi-AZ machinesets for compute.
Not that I know of, but it's used by the AWS actuator to determine if a machine already exists:
It's only needed for the AWS actuator to match an existing instance with a machine object. For libvirt, the machine name is sufficient (https://github.com/openshift/cluster-api-provider-libvirt/blob/master/cloud/libvirt/actuators/machine/utils/domain.go#L617-L633). For OpenStack, the name, image + flavor are used (https://github.com/openshift/cluster-api-provider-openstack/blob/master/cloud/openstack/machineactuator.go#L434-L452) |
👍 |
Error: Error applying plan:
6 error(s) occurred:
* module.vpc.aws_eip.nat_eip[1]: 1 error(s) occurred:
* aws_eip.nat_eip.1: Error creating EIP: AddressLimitExceeded: The maximum number of addresses has been reached.
status code: 400, request id: 68d2f8d1-521f-45b9-8a01-1f5814bd79cf
* module.vpc.aws_eip.nat_eip[5]: 1 error(s) occurred:
* aws_eip.nat_eip.5: Error creating EIP: AddressLimitExceeded: The maximum number of addresses has been reached.
status code: 400, request id: d32497d3-b5fb-42cc-8b15-711417c27cb4
* module.vpc.aws_eip.nat_eip[0]: 1 error(s) occurred:
* aws_eip.nat_eip.0: Error creating EIP: AddressLimitExceeded: The maximum number of addresses has been reached.
status code: 400, request id: 94ece091-7cd6-49a5-9377-fa5a4bf7c19e
* module.vpc.aws_eip.nat_eip[3]: 1 error(s) occurred:
* aws_eip.nat_eip.3: Error creating EIP: AddressLimitExceeded: The maximum number of addresses has been reached.
status code: 400, request id: 88ed257a-abfc-4420-b40a-9b1863e73a92
* module.vpc.aws_eip.nat_eip[2]: 1 error(s) occurred:
* aws_eip.nat_eip.2: Error creating EIP: AddressLimitExceeded: The maximum number of addresses has been reached.
status code: 400, request id: b31bc1ab-2680-4430-b3ad-5e3a58980468
* module.vpc.aws_eip.nat_eip[4]: 1 error(s) occurred:
* aws_eip.nat_eip.4: Error creating EIP: AddressLimitExceeded: The maximum number of addresses has been reached.
status code: 400, request id: 1e76b769-fd09-4883-be84-d13e149f51a1
Terraform does not automatically rollback in the face of errors.
Instead, your Terraform state file has been partially updated with
any resources that successfully completed. Please address the error
above and apply again to incrementally change your infrastructureHopefully CI cleared up fingers crossed |
|
@csrwng @abhinavdahiya I don't think there's docs on that tag yet, will fix that.
My understanding is that tf assign each master a subnet https://github.com/openshift/installer/blob/master/data/data/aws/master/main.tf#L83 |
|
/retest |
|
/test all |
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: abhinavdahiya, csrwng 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 |
openshift-ci-robot
added
the
approved
Closing the loop on this, here is @csrwng's link from this comment pinned to a particular commit (vs. floating with master). And the difference is that The cluster-API AWS provider has also been setting And we are setting the same |
Centralize extra-tag inclusion on aws/main.tf. This reduces the
number of places we need to think about what tags should be ;).
Also keep kubernetes.io/cluster/{name} localized in the aws module.
See 8a37f72 (modules/aws/bootstrap: Pull AWS bootstrap setup into a
module, 2018-09-05, openshift#217) for why we need to keep it on the bootstrap
instance. But the bootstrap resources will be removed after the
bootstrap-complete event comes through, and we don't want Kubernetes
controllers trying to pick them up.
This commit updates the internal Route 53 zone from KubernetesCluster
to kubernetes.io/cluster/{name}: owned, catching it up to
kubernetes/kubernetes@0b5ae539 (AWS: Support shared tag, 2017-02-18,
kubernetes/kubernetes#41695). That tag originally landed on the zone
back in 75fb49a (platforms/aws: apply tags to internal route53 zone,
2017-05-02, coreos/tectonic-installer#465).
Only the master instances need the clusterid tag, as described in
6c7a5f0 (Tag master machines for adoption by machine controller,
2018-10-17, openshift#479).
A number of VPC resources have moved from "shared" to "owned". The
shared values are from 45dfc2b (modules/aws,azure: use the new tag
format for k8s 1.6, 2017-05-04, coreos/tectonic-installer#469). The
commit message doesn't have much to say for motivation, but Brad Ison
said [1]:
I'm not really sure if anything in Kubernetes actually uses the
owned vs. shared values at the moment, but in any case, it might
make more sense to mark subnets as shared. That was actually one of
the main use cases for moving to this style of tagging -- being able
to share subnets between clusters.
But we aren't sharing these resources; see 6f55e67 (terraform/aws:
remove option to use an existing vpc in aws, 2018-11-11, openshift#654).
[1]: coreos/tectonic-installer#469 (comment)
Centralize extra-tag inclusion on aws/main.tf. This reduces the
number of places we need to think about what tags should be ;).
Also keep kubernetes.io/cluster/{name} localized in the aws module.
See 8a37f72 (modules/aws/bootstrap: Pull AWS bootstrap setup into a
module, 2018-09-05, openshift#217) for why we need to keep it on the bootstrap
instance. But the bootstrap resources will be removed after the
bootstrap-complete event comes through, and we don't want Kubernetes
controllers trying to pick them up.
This commit updates the internal Route 53 zone from KubernetesCluster
to kubernetes.io/cluster/{name}: owned, catching it up to
kubernetes/kubernetes@0b5ae539 (AWS: Support shared tag, 2017-02-18,
kubernetes/kubernetes#41695). That tag originally landed on the zone
back in 75fb49a (platforms/aws: apply tags to internal route53 zone,
2017-05-02, coreos/tectonic-installer#465).
Only the master instances need the clusterid tag, as described in
6c7a5f0 (Tag master machines for adoption by machine controller,
2018-10-17, openshift#479).
A number of VPC resources have moved from "shared" to "owned". The
shared values are from 45dfc2b (modules/aws,azure: use the new tag
format for k8s 1.6, 2017-05-04, coreos/tectonic-installer#469). The
commit message doesn't have much to say for motivation, but Brad Ison
said [1]:
I'm not really sure if anything in Kubernetes actually uses the
owned vs. shared values at the moment, but in any case, it might
make more sense to mark subnets as shared. That was actually one of
the main use cases for moving to this style of tagging -- being able
to share subnets between clusters.
But we aren't sharing these resources; see 6f55e67 (terraform/aws:
remove option to use an existing vpc in aws, 2018-11-11, openshift#654).
[1]: coreos/tectonic-installer#469 (comment)
Centralize extra-tag inclusion on aws/main.tf. This reduces the
number of places we need to think about what tags should be ;).
Also keep kubernetes.io/cluster/{name} localized in the aws module.
See 8a37f72 (modules/aws/bootstrap: Pull AWS bootstrap setup into a
module, 2018-09-05, openshift#217) for why we need to keep it on the bootstrap
instance. But the bootstrap resources will be removed after the
bootstrap-complete event comes through, and we don't want Kubernetes
controllers trying to pick them up.
This commit updates the internal Route 53 zone from KubernetesCluster
to kubernetes.io/cluster/{name}: owned, catching it up to
kubernetes/kubernetes@0b5ae539 (AWS: Support shared tag, 2017-02-18,
kubernetes/kubernetes#41695). That tag originally landed on the zone
back in 75fb49a (platforms/aws: apply tags to internal route53 zone,
2017-05-02, coreos/tectonic-installer#465).
Only the master instances need the clusterid tag, as described in
6c7a5f0 (Tag master machines for adoption by machine controller,
2018-10-17, openshift#479).
A number of VPC resources have moved from "shared" to "owned". The
shared values are from 45dfc2b (modules/aws,azure: use the new tag
format for k8s 1.6, 2017-05-04, coreos/tectonic-installer#469). The
commit message doesn't have much to say for motivation, but Brad Ison
said [1]:
I'm not really sure if anything in Kubernetes actually uses the
owned vs. shared values at the moment, but in any case, it might
make more sense to mark subnets as shared. That was actually one of
the main use cases for moving to this style of tagging -- being able
to share subnets between clusters.
But we aren't sharing these resources; see 6f55e67 (terraform/aws:
remove option to use an existing vpc in aws, 2018-11-11, openshift#654).
[1]: coreos/tectonic-installer#469 (comment)
Add a
clusteridtag to masters so they can be adopted by the machine controller in the future.