Automatic merge from submit-queue.
Allow additional SGs to be added to API loadbalancer
Allow adding precreated additional security groups to the API loadbalancer using cluster spec:
```yaml
spec:
api:
loadBalancer:
type: Public
additionalSecurityGroups:
- sg-exampleid3
- sg-exampleid4
```
- [x] Adding additionalSecurityGroups cluster spec
- [x] Adding validation for repeated security groups
- [x] Adding validation for API loadbalancer security groups
- [x] Integration test for API loadbalancer and its security groups
- [x] Update API docs and cluster.spec docs
Automatic merge from submit-queue.
Add additionalNetworkCIDRs to support VPCs with multiple CIDRs in AWS
Add additionalNetworkCIDRs to support VPCs with multiple CIDRs in AWS.
@justinsb I cannot find anywhere that does a check on an existing VPC to see if the networkCIDR matches what is on the VPC defined, I was looking for that so I can add a similar check for this. Am I missing something or is there really no check like that?
Automatic merge from submit-queue.
Fix spurious shared InternetGateway renaming
This fixes the original issue described in #476 where `kops update cluster` would continuously report renaming a shared IGW even though it never actually renames it. I also added tests to confirm the behavior change.
I removed the bug mention from the docs although if you would prefer to keep it and instead say `In kops versions before X, ...` I can do that too
Automatic merge from submit-queue.
Minor update to docs/aws.md
Updates the backtick syntax to `$(...)`, plus a minor grammar fix.
I wasn't sure if I needed to fill out a CLA for just a documentation update. Let me know if I should!
_Update:_ Never mind, I've signed the agreement now!
Automatic merge from submit-queue.
Add edit flag so create instancegroup command is usable on scripts
Until now, the opening of an interactive editor when creating a new instance group was mandatory.
This this commit, a new flag is added, so this is now optional.
This commit is backwards compatible.
* Limit each CNI provider to 100m
* Remove CPU limits - they cause serious problems
(https://github.com/kubernetes/kubernetes/issues/51135), but this also
makes the CPU allocation less problematic.
* Bump versions and start introducing the `-kops.1` suffix preemptively.
* Upgrade flannel to 0.9.0 as it fixes a lot.
Automatic merge from submit-queue.
Consolidate two separate CNI sections
There was a second section about CNI that was confusing since it was superfluous to the first one.
Also updated a couple of calico references from version-pinned to 'latest'.