fix(addons/networking.projectcalico.org) calico kube-controllers is needed in CRD mode

[phspagiari]

The code comment says that Calico Kube Controllers isn't need in CRD mode, but we can see docs of calico saying how to configure for datastore mode and we started having problems using calico without this.

Problem description

• We start seeing logs like this in calico-node:

2019-09-03 14:50:41.821 [INFO][24] ipam.go 311: Looking up existing affinities for host handle="ipip-tunnel-addr-ip-10-164-41-80.ec2.internal" host="ip-10-164-41-80.ec2.internal"
2019-09-03 14:50:41.966 [INFO][24] ipam.go 369: Ran out of existing affine blocks for host handle="ipip-tunnel-addr-ip-10-164-41-80.ec2.internal" host="ip-10-164-41-80.ec2.internal"
2019-09-03 14:50:41.971 [INFO][24] ipam.go 434: No more affine blocks, but need to allocate 1 more addresses - allocate another block handle="ipip-tunnel-addr-ip-10-164-41-80.ec2.internal" host="ip-10-164-41-80.ec2.internal"
2019-09-03 14:50:41.971 [INFO][24] ipam.go 438: Looking for an unclaimed block handle="ipip-tunnel-addr-ip-10-164-41-80.ec2.internal" host="ip-10-164-41-80.ec2.internal"

• New nodes stopped to be up and Estabilished in calicoctl node status
• We can see a lot of unused ipamblocks and blockaffinities resources on K8S with a very different number when compared to our nodes running:

kubectl get ipamblocks | wc -l && kubectl get blockaffinities | wc -l && kubectl get nodes | wc -l
512
512
76

• After applying the deployment of calico-kube-controller, everything started working again and the numbers are now correct:

kubectl get ipamblocks | wc -l && kubectl get blockaffinities | wc -l && kubectl get nodes | wc -l
76
76
76

• We can see calico-kube-controllers logs doing what was expected here:

2019-09-04 13:16:09.681 [INFO][1] ipam.go 1166: Releasing all IPs with handle 'ipip-tunnel-addr-ip-10-164-76-215.ec2.internal'
2019-09-04 13:16:13.677 [INFO][1] kdd.go 154: Cleaning up IPAM resources for deleted node node="ip-10-164-41-56.ec2.internal"

[k8s-ci-robot]

Hi @phspagiari. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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.

[mikesplain]

/ok-to-test
/assign

[mikesplain]

That's a good catch, looks like there used to be old documentation stating this but it's been changed for current versions: https://github.com/projectcalico/kube-controllers/issues/346#issuecomment-470313052

[mikesplain]

/retest

[mikesplain]

Tested this myself as well. We probably want to do some updates soon but lets get this back in now. I'm going to open a cherry pick to 1.13 and 1.14 too.

Thanks so much!
/lgtm
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: mikesplain, phspagiari

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:

• OWNERS [mikesplain]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[mikesplain]

Just to document the cherry picks of this better:

#7517 => original PR (master/v1.15)
#7526 => CP of #7526 to 1.13 modified with older versions
#7527 => CP of #7526 to 1.14 with same versions