Commit Graph

14 Commits

Author SHA1 Message Date
Dalton Hubble d774c51297 Update Kubernetes from v1.8.2 to v1.8.3 2017-11-08 23:34:19 -08:00
Dalton Hubble f6a8fb363e Remove deprecated kubelet --require-kubeconfig flag
* https://github.com/kubernetes/kubernetes/pull/40050
2017-11-08 23:34:19 -08:00
Dalton Hubble 451fd86470 Improve internal firewall rules on Google Cloud
* Whitelist internal traffic between controllers and workers
* Switch to tag-based firewall policies rather than source IP
2017-11-08 00:15:06 -08:00
Dalton Hubble 168c487484 Remove mention of self-hosted etcd, its deprecated 2017-11-06 01:03:53 -08:00
Dalton Hubble 2b0296d671 Create controller instances across zones in the region
* Change controller instances to automatically span zones in a region
* Remove the `zone` required variable
2017-11-05 13:24:32 -08:00
Dalton Hubble 7b38271212 Run etcd cluster on-host, across controllers on Google Cloud
* Change controllers from a managed group to individual instances
* Create discrete DNS records to each controller's private IP for etcd
* Change etcd to run on-host, across controllers (etcd-member.service)
* Reduce time to bootstrap a cluster
* Deprecate self-hosted-etcd on the Google Cloud platform
2017-11-05 11:03:35 -08:00
Dalton Hubble a48dd9ebd8 Require google provider version ~> 1.1
* Require google provider plugin 1.1 or higher which includes fix:
https://github.com/terraform-providers/terraform-provider-google/issues/574
* Remove workaround which statically set the persistent disk name
* Original reasons for workaround in a97df839 or GH #34
2017-11-04 12:59:19 -07:00
Dalton Hubble 6300383b43 Change worker managed instance group to span zones in region
* Change Google Cloud module to require the `region` variable
* Workers are created in random zones within the given region
* Tolerate Google Cloud zone failures or capacity issues
* If workers are preempted (if enabled), replacement instances can
be drawn from any zone in the region, which should avoid scheduling
issues that were possible before if a single zone aggressively
preempts instances (presumably due to Google Cloud capacity)
2017-11-04 12:59:19 -07:00
Dalton Hubble 60bc8957c9 Update Kubernetes from v1.8.1 to v1.8.2
* Kubernetes v1.8.2 fixes a memory leak in the v1.8.1 apiserver
* Switch to using the `gcr.io/google_containers/hyperkube` for the
on-host kubelet and shutdown drains
* Update terraform-render-bootkube manifests generation
  * Update flannel from v0.8.0 to v0.9.0
  * Add `hairpinMode` to flannel CNI config
  * Add `--no-negcache` to kube-dns dnsmasq
2017-10-24 21:44:26 -07:00
Dalton Hubble 8b78c65483 Update Google Cloud Kubernetes from v1.7.7 to v1.8.1 2017-10-20 16:09:11 -07:00
Dalton Hubble 43dc44623f Fix the terraform fmt of configs 2017-10-16 01:32:25 -07:00
Dalton Hubble a97df839ea google-cloud: Set disk.device_name to match API default
* Terraform provider "google" plugin releases leave the disk
device_name as "" by default. Recently the API has started to
set a default name "persistent-disk-0". Plan and apply show
all instance groups need to be recreated to "fix" the name
* Impact: Controller and worker instance groups are deleted
and recreated, deleting data on controllers and bringing
down clusters
* Fix: Explicitly set the disk_name to persistent-disk-0 so
that terraform finds no diff needs to be applied.
* https://github.com/poseidon/typhoon/issues/34
* https://github.com/terraform-providers/terraform-provider-google/issues/574
2017-10-11 18:04:39 -07:00
Dalton Hubble 1bc25c1036 Update Kubernetes from v1.7.5 to v1.7.7
* Update from bootkube v0.6.2 to v0.7.0
* Use renamed terraform-render-bootkube. Renamed from
bootkube-terraform to meet Terraform Module requirements
2017-10-03 21:03:15 -07:00
Dalton Hubble 795428329a google-cloud: Move controller and worker submodules under kubernetes 2017-09-27 20:50:32 -07:00