Commit graph

347 commits

Author SHA1 Message Date
Robert Everson
ac18f6cf8b Add support for admission controllers in 1.10 and above 2018-08-07 11:25:03 -07:00
Rong Zhang
c288ffc55d
Merge pull request #2342 from southquist/add-ca-cert
allow for setting the cacert on openstack cloud provider
2018-08-07 17:46:01 +08:00
Aivars Sterns
72f053d9bb
Merge pull request #2972 from mattymo/force_cni_cp
Force copy cni files
2018-07-10 09:40:10 +03:00
Dao Hoang Son
d306c9708c Remove step that force disable kube_basic_auth.
The referenced issue (https://github.com/kubernetes/kubeadm/issues/441) has already been fixed.
2018-07-08 16:57:43 +07:00
Matthew Mosesohn
1a3b9dd864 Force copy cni files 2018-07-06 16:39:42 +03:00
Miouge1
2a279e30b0 CheckNodePIDPressure is not supported in v1.10 2018-06-28 20:10:38 +02:00
southquist
c685dc493f allow for setting the cacert on openstack cloud provider 2018-06-28 16:00:13 +02:00
Andreas Krüger
cbb959151c
Merge pull request #2737 from Miouge1/update-scheduler
Update kube-scheduler policy
2018-06-19 14:53:22 +02:00
Matthew Mosesohn
61e97251a5 Improve variable handling for disabling etcd events cluster 2018-06-18 16:58:29 +03:00
Andreas Krüger
e60a63ea51
Merge pull request #2577 from woopstar/etcd-fix-4
Makeover of etcd- and etcd-cluster setup.
2018-05-16 20:49:54 +02:00
Matthew Mosesohn
7c93e71801
Upgrade k8s to 1.10.2 (#2748)
* Upgrade k8s to 1.10.2

Bumped etcd version to 3.2.16 as recommended

* Add ipvs fix for v1.10

* change flannel addons test to ha
2018-05-15 16:00:29 +03:00
Christopher J. Ruwe
73800ef111 make certificates non-executable 2018-05-15 07:54:32 +00:00
Miouge1
ad48606e4e Restart scheduler when policy changes 2018-05-14 10:09:30 +02:00
Matthew Mosesohn
07cc981971
refactor vault role (#2733)
* Move front-proxy-client certs back to kube mount

We want the same CA for all k8s certs

* Refactor vault to use a third party module

The module adds idempotency and reduces some of the repetitive
logic in the vault role

Requires ansible-modules-hashivault on ansible node and hvac
on the vault hosts themselves

Add upgrade test scenario
Remove bootstrap-os tags from tasks

* fix upgrade issues

* improve unseal logic

* specify ca and fix etcd check

* Fix initialization check

bump machine size
2018-05-11 19:11:38 +03:00
Andreas Krüger
28d6eb6af1
Merge pull request #2644 from cp3hu/master
Fix apiserver manifest and kubelet for kube version < 1.9
2018-05-08 09:22:36 +02:00
Miouge1
70e0998a70 Update kube-scheduler policy 2018-05-03 21:56:51 +02:00
woopstar
4c81cd2a71 Merge branch 'master' of https://github.com/kubernetes-incubator/kubespray into etcd-fix-4 2018-05-02 14:45:58 +02:00
Suzuka Asagiri
f81e6d2ccf
Add oidc-user-prefix and oidc-group-prefix args 2018-04-23 12:23:59 +09:00
Chad Swenson
d87b6fd9f3 Use dedicated front-proxy-ca for front-proxy-client 2018-04-12 11:03:22 -05:00
Christian Phu
3535c29e59 Fix apiserver manifest for kube version < 1.9 2018-04-10 18:17:56 +02:00
Marcelo Grebois
88765f62e6
Updating order
https://kubernetes.io/docs/admin/admission-controllers/#is-there-a-recommended-set-of-admission-controllers-to-use
2018-04-10 17:17:39 +02:00
Marcelo Grebois
4c12b273ac
Enabling MutatingAdmissionWebhook for Istio Automatic sidecar injection
https://istio.io/docs/setup/kubernetes/sidecar-injection.html#automatic-sidecar-injection
2018-04-09 12:49:05 +02:00
Andreas Krüger
2511e14289
Merge pull request #2346 from Miouge1/kube-scheduler-mode
Use legacy policy config to apply the scheduler policy
2018-04-04 10:20:51 +02:00
woopstar
86e3506ae6 Etcd cluster setup makeover
The current way to setup the etc cluster is messy and buggy.

- It checks for cluster is healthy before the cluster is even created.
- The unit files are started on handlers, not in the task, so you mess with "flush handlers".
- The join_member.yml is not used.
- etcd events cluster is not configured for kubeadm
- remove duplicate runs between running the role on etcd nodes and k8s nodes
2018-04-01 21:38:33 +02:00
Wong Hoi Sing Edison
195d6d791a Integrate jetstack/cert-manager 0.2.3 to Kubespray 2018-03-31 19:29:11 +08:00
Andreas Krüger
d9418b1dc4
Merge pull request #2554 from georgejdli/fix-sa-token-signing
Fix kubespray's ServiceAccount token signing keys
2018-03-31 09:59:22 +02:00
Andreas Krüger
76cb37d6b5
Merge pull request #2544 from woopstar/cert-fix-2
Update openssl.conf to count better and work with Jinja 2.9
2018-03-30 21:57:17 +02:00
georgejdli
572ab650db copy dedicated service account token signing key for kubeadm migration 2018-03-30 13:03:32 -05:00
Matthew Mosesohn
03bcfa7ff5
Stop templating kube-system namespace and creating it (#2545)
Kubernetes makes this namespace automatically, so there is
no need for kubespray to manage it.
2018-03-30 14:29:13 +03:00
Andreas Kruger
af5f376163 Revert 2018-03-30 11:42:20 +02:00
woopstar
004b0a3fcf Fix merge conflict 2018-03-30 11:38:59 +02:00
Andreas Krüger
f619eb08b1
Merge pull request #2350 from whereismyjetpack/kubeadm-nodename
set nodeName to "{{ inventory_hostname }}" in kubeadm-config
2018-03-30 11:15:52 +02:00
Kuldip Madnani
daeeae1a91 Added retries in pre-upgrade.yml and retries while applying kube-dns.yml (#2553)
* Added retries in pre-upgrade.yml and retries while applying kube-dns.yml

* Removed trailing spaces
2018-03-29 11:37:32 -05:00
georgejdli
c8f857eae4 configure kubespray to sign service account tokens with a dedicated and stable key 2018-03-29 09:50:31 -05:00
Dann Bohn
1d0415a6cf fixes typo in kube_override_hostname for kubeadm 2018-03-24 13:29:07 -04:00
Dann Bohn
9fa995ac9d only sets nodeName in kubeadm-config when kube_override_hostname is set 2018-03-23 08:33:25 -04:00
Andreas Krüger
30e4b89837
Merge pull request #2504 from brtknr/patch-1
Update kube-apiserver.manifest.j2 and kubeadm-config.yaml.j2 to incorporate `endpoint-reconciler-type: lease`
2018-03-22 09:15:55 +01:00
Andreas Krüger
ff2b8e5e60
Merge pull request #2503 from woopstar/kubelet-fix-1
Fix duplicate --proxy-client-cert-file and --proxy-client-key-file
2018-03-21 10:03:31 +01:00
mirwan
ee8f678010 Addition of the .creds extension to the credentials files generated by password lookup in order for Ansible not to consider them as inventory files with inventory_ignore_extensions set accordingly (#2446) 2018-03-21 10:50:32 +03:00
Bharat Kunwar
13e47e73c8
Update kubeadm-config.yaml.j2
As requested
2018-03-20 13:33:36 +00:00
Bharat Kunwar
d2fd7b7462
Update kube-apiserver.manifest.j2 2018-03-20 12:19:53 +00:00
Bharat Kunwar
d9453f323b
Update kube-apiserver.manifest.j2 2018-03-20 12:16:35 +00:00
Bharat Kunwar
b787b76c6c
Update kube-apiserver.manifest.j2
Ensure that kube-apiserver will respond even if one of the nodes are down.
2018-03-20 12:06:34 +00:00
woopstar
a94a407a43 Fix duplicate --proxy-client-cert-file and --proxy-client-key-file 2018-03-20 12:08:36 +01:00
Andreas Krüger
d29a1db134
Merge pull request #2461 from woopstar/patch-11
Add support to kubeadm too
2018-03-16 08:24:31 +01:00
Andreas Krüger
653d97dda4
Merge pull request #2472 from woopstar/patch-12
Make sure output from extra args is strings
2018-03-16 08:23:50 +01:00
woopstar
40c0f3756b Encapsulate item instead of casting to string 2018-03-15 20:27:21 +01:00
Andreas Krüger
3d6fd49179 Added option for encrypting secrets to etcd v.2 (#2428)
* Added option for encrypting secrets to etcd

* Fix keylength to 32

* Forgot the default

* Rename secrets.yaml to secrets_encryption.yaml

* Fix static path for secrets file to use ansible variable

* Rename secrets.yaml.j2 to secrets_encryption.yaml.j2

* Base64 encode the token

* Fixed merge error

* Changed path to credentials dir

* Update path to secrets file which is now readable inside the apiserver container. Set better file permissions

* Add encryption option to k8s-cluster.yml
2018-03-15 22:20:05 +03:00
Andreas Krüger
788e41a315
Make sure output from extra args is strings
Setting the following:

```
kube_kubeadm_controller_extra_args:
  address: 0.0.0.0
  terminated-pod-gc-threshold: "100"
```

Results in `terminated-pod-gc-threshold: 100` in the kubeadm config file. But it has to be a string to work.
2018-03-14 19:23:43 +01:00
Andreas Krüger
39d247a238
Add support to kubeadm too
Explicitly defines the --kubelet-preferred-address-types parameter #2418

Fixes #2453
2018-03-13 10:31:15 +01:00
Ayaz Ahmed Khan
89847d5684 Explicitly defines the --kubelet-preferred-address-types parameter
to the API server configuration.

This solves the problem where if you have non-resolvable node names,
and try to scale the server by adding new nodes, kubectl commands
start to fail for newly added nodes, giving a TCP timeout error when
trying to resolve the node hostname against a public DNS.
2018-03-05 15:25:14 +01:00
RongZhang
67ffd8e923 Add etcd-events cluster for kube-apiserver (#2385)
Add etcd-events cluster for kube-apiserver
2018-03-01 11:39:14 +03:00
Nedim Haveric
2bd3776ddb fix apiserver manifest when disabling insecure_port 2018-02-22 14:00:32 +01:00
Andreas Krüger
d84ff06f73 Set filemode to 0640 (#2315)
* Set filemode to 0640

weave-net.yml file is readable by all users on the host. It however contains the weave_password to encrypt all pod communication. It should only be readable by root.

* Set mode 0640 on users_file with basic auth
2018-02-21 23:13:46 +03:00
melkosoft
f13e76d022 Added cilium support (#2236)
* Added cilium support

* Fix typo in debian test config

* Remove empty lines

* Changed cilium version from <latest> to <v1.0.0-rc3>

* Add missing changes for cilium

* Add cilium to CI pipeline

* Fix wrong file name

* Check kernel version for cilium

* fixed ci error

* fixed cilium-ds.j2 template

* added waiting for cilium pods to run

* Fixed missing EOF

* Fixed trailing spaces

* Fixed trailing spaces

* Fixed trailing spaces

* Fixed too many blank lines

* Updated tolerations,annotations in cilium DS template

* Set cilium_version to iptables-1.9 to see if bug is fixed in CI

* Update cilium image tag to v1.0.0-rc4

* Update Cilium test case CI vars filenames

* Add optional prometheus flag, adjust initial readiness delay

* Update README.md with cilium info
2018-02-16 21:37:47 -06:00
Dann Bohn
95e2bde15b set nodeName to "{{ inventory_hostname }}" in kubeadm-config 2018-02-16 16:20:08 -05:00
Miouge1
4c280e59d4 Use legacy policy config to apply the scheduler policy 2018-02-16 13:43:35 +01:00
Virgil Chereches
d72232f15b Increased timeout values for k8s API server restart 2018-02-12 07:35:29 +00:00
Maxim Krasilnikov
03c61685fb
Added apiserver extra args variable for kubeadm config (#2291) 2018-02-12 10:29:46 +03:00
Chia-liang Kao
338238d086
Fix version comparison
`FAILED! => {"changed": false, "msg": "AnsibleFilterError: Version comparison: unorderable types: str() < int()"}`
2018-02-10 03:49:49 +08:00
mlushpenko
4e61fb9cd3 Refactored kubeadm join process and fixed uncrodonng for master nodes 2018-02-09 15:51:47 +01:00
Antoine Legrand
7bce70339f
Merge pull request #2251 from woopstar/metrics-server-patch-2
Adding metrics-server support for K8s version 1.9
2018-02-08 11:16:44 +01:00
woopstar
f193b12059 Kubeadm auto creates this 2018-02-07 10:50:34 +01:00
woopstar
2cd254954c Remove defaults of allowed names. Updated kubeadm 2018-02-07 10:07:55 +01:00
woopstar
4dab92ce69 Rename from aggregator-proxy-client to front-proxy-client to match kubeadm design. Added kubeadm support too. Changed to use variables set and not hardcode paths. Still missing cert generation for Vault 2018-02-07 09:50:19 +01:00
woopstar
b2d30d68e7 Rename CN for aggreator back. Add flags to apiserver when version is >= 1.9 2018-02-05 20:37:14 +01:00
Maxim Krasilnikov
95b8ac5f62 Added optional controller and scheduler extra args to kubeadm config (#2205) 2018-02-05 16:49:13 +03:00
Chad Swenson
bd1f0bcfd7
Merge pull request #2201 from riverzhang/ipvs
Support ipvs mode for kube-proxy
2018-02-01 22:29:52 -06:00
Dann Bohn
dc6c703741 --etcd-quorum-read is depricated in kube >= 1.9 2018-01-31 15:49:52 -05:00
rong.zhang
b10c308a5a Support ipvs mode for kube-proxy
Support ipvs mode for kube-proxy
2018-01-30 13:09:01 +08:00
Matthew Mosesohn
dc6a17e092
Use include/import tasks (#2192)
import_tasks will consume far less memory, so it should be
used whenever it is compatible.
2018-01-29 14:37:48 +03:00
Chad Swenson
c6e0fcea31
Merge pull request #1948 from sgmitchell/secured-etcd
Enable etcd secure client to prevent etcdctl access without cert and key
2018-01-25 09:35:51 -06:00
Brad Beam
0c8bed21ee
Merge pull request #2019 from chadswen/disable-api-insecure-port
Support for disabling apiserver insecure port (the sequel)
2018-01-24 19:58:53 -06:00
Brad Beam
98300e3165
Merge pull request #2155 from brutus333/fix/pvc
Fix for Issue #2141
2018-01-24 16:15:33 -06:00
Virgil Chereches
a4d142368b Renamed variable from disable_volume_zone_conflict to volume_cross_zone_attachment and removed cloud provider condition; fix identation 2018-01-23 13:14:00 +00:00
Virgil Chereches
3125f93b3f Added disable_volume_zone_conflict variable 2018-01-18 10:55:23 +00:00
Virgil Chereches
8c45c88d15 Fix for Issue #2141 - added policy file 2018-01-12 07:15:35 +00:00
Virgil Chereches
c87bb2f239 Fix for Issue #2141 2018-01-12 07:07:02 +00:00
Spencer Smith
ef96123482
Merge pull request #2068 from chadswen/remove-container-retries
Retry kube container removal during upgrade
2018-01-09 15:03:50 -05:00
Steve Mitchell
e45b30d033 Add etcd key and cert environment variables for use with client auth 2018-01-02 13:52:17 -05:00
Matthew Mosesohn
ad6fecefa8
Update Kubernetes to v1.9.0 (#2100)
Update checksum for kubeadm
Use v1.9.0 kubeadm params
Include hash of ca.crt for kubeadm join
Update tag for testing upgrades
Add workaround for testing upgrades
Remove scale CI scenarios because of slow inventory parsing
in ansible 2.4.x.

Change region for tests to us-central1 to
improve ansible performance
2017-12-25 08:57:45 +00:00
Matthew Mosesohn
6bb46e3ecb
Fix param names in preparation for Kubernetes v1.9.0 (#2098)
This does not update v1.9.0, but fixes two incompatibilities
when trying to deploy v1.9.0.
2017-12-20 10:48:09 +00:00
Chad Swenson
e78562830f Retry kube container removal during upgrade
As we have seen with other containers, sometimes container removal fails on the first attempt due to some Docker bugs. Retrying typically corrects the issue.
2017-12-12 12:06:41 -06:00
Spencer Smith
c4458c9d9a
Merge pull request #1997 from mrbobbytables/feature-keepalived-cloud-provider
Add minimal keepalived-cloud-provider support
2017-12-06 23:28:27 -05:00
Chad Swenson
b8788421d5 Support for disabling apiserver insecure port
This allows `kube_apiserver_insecure_port` to be set to 0 (disabled).

Rework of #1937 with kubeadm support

Also, fixed an issue in `kubeadm-migrate-certs` where the old apiserver cert was copied as the kubeadm key
2017-12-05 09:13:45 -06:00
Steven Hardy
d39a88d63f Allow setting --bind-address for apiserver hyperkube (#1985)
* Allow setting --bind-address for apiserver hyperkube

This is required if you wish to configure a loadbalancer (e.g haproxy)
running on the master nodes without choosing a different port for the
vip from that used by the API - in this case you need the API to bind to
a specific interface, then haproxy can bind the same port on the VIP:

root@overcloud-controller-0 ~]# netstat -taupen | grep 6443
tcp        0      0 192.168.24.6:6443       0.0.0.0:*               LISTEN      0          680613     134504/haproxy
tcp        0      0 192.168.24.16:6443      0.0.0.0:*               LISTEN      0          653329     131423/hyperkube
tcp        0      0 192.168.24.16:6443      192.168.24.16:58404     ESTABLISHED 0          652991     131423/hyperkube
tcp        0      0 192.168.24.16:58404     192.168.24.16:6443      ESTABLISHED 0          652986     131423/hyperkube

This can be achieved e.g via:

kube_apiserver_bind_address: 192.168.24.16

* Address code review feedback

* Update kube-apiserver.manifest.j2
2017-11-29 15:24:02 +00:00
Spencer Smith
43600ffcf8
Merge pull request #1972 from chadswen/master-static-pod-flush
Additional flush for static pod master upgrade
2017-11-28 11:36:38 -05:00
Bogdan Dobrelya
8aafe64397
Defaults for apiserver_loadbalancer_domain_name (#1993)
* Defaults for apiserver_loadbalancer_domain_name

When loadbalancer_apiserver is defined, use the
apiserver_loadbalancer_domain_name with a given default value.

Fix unconsistencies for checking if apiserver_loadbalancer_domain_name
is defined AND using it with a default value provided at once.

Signed-off-by: Bogdan Dobrelya <bogdando@mail.ru>

* Define defaults for LB modes in common defaults

Adjust the defaults for apiserver_loadbalancer_domain_name and
loadbalancer_apiserver_localhost to come from a single source, which is
kubespray-defaults. Removes some confusion and simplefies the code.

Signed-off-by: Bogdan Dobrelya <bogdando@mail.ru>
2017-11-23 16:15:48 +00:00
Bob Killen
2140303fcc
add minimal keepalived-cloud-provider support 2017-11-23 08:43:36 -05:00
Matthew Mosesohn
f9b68a5d17
Revert "Support for disabling apiserver insecure port" (#1974) 2017-11-14 13:41:28 +00:00
Chad Swenson
1f99710b21 Additional flush for static pod master upgrade
Thought this wasn't required at first but I forgot there's no auto flush at the end of these tasks since the `kubernetes/master` role is not the end of the play.
2017-11-13 18:11:57 -06:00
Spencer Smith
09d85631dc
Merge pull request #1944 from chadswen/reload-master-pods
Master component and kubelet container upgrade fixes
2017-11-08 22:23:12 -05:00
Chad Swenson
e9f795c5ce Master component and kubelet container upgrade fixes
* Fixes an issue where apiserver and friends (controller manager, scheduler) were prevented from restarting after manifests/secrets are changed. This occurred when a replaced kubelet doesn't reconcile new master manifests, which caused old master component versions to linger during deployment. In my case this was causing upgrades from k8s 1.6/1.7 -> k8s 1.8 to fail
* Improves transitions from kubelet container to host kubelet by preventing issues where kubelet container reappeared during the deployment
2017-11-08 01:40:33 -06:00
Chad Swenson
0c7e1889e4 Support for disabling apiserver insecure port
This allows `kube_apiserver_insecure_port` to be set to 0 (disabled). It's working, but so far I have had to:

1. Make the `uri` module "Wait for apiserver up" checks use `kube_apiserver_port` (HTTPS)
2. Add apiserver client cert/key to the "Wait for apiserver up" checks
3. Update apiserver liveness probe to use HTTPS ports
4. Set `kube_api_anonymous_auth` to true to allow liveness probe to hit apiserver's /healthz over HTTPS (livenessProbes can't use client cert/key unfortunately)
5. RBAC has to be enabled. Anonymous requests are in the `system:unauthenticated` group which is granted access to /healthz by one of RBAC's default ClusterRoleBindings. An equivalent ABAC rule could allow this as well.

Changes 1 and 2 should work for everyone, but 3, 4, and 5 require new coupling of currently independent configuration settings. So I also added a new settings check.

Options:

1. The problem goes away if you have both anonymous-auth and RBAC enabled. This is how kubeadm does it. This may be the best way to go since RBAC is already on by default but anonymous auth is not.
2. Include conditional templates to set a different liveness probe for possible combinations of `kube_apiserver_insecure_port = 0`, RBAC, and `kube_api_anonymous_auth` (won't be possible to cover every case without a guaranteed authorizer for the secure port)
3. Use basic auth headers for the liveness probe (I really don't like this, it adds a new dependency on basic auth which I'd also like to leave independently configurable, and it requires encoded passwords in the apiserver manifest)

Option 1 seems like the clear winner to me, but is there a reason we wouldn't want anonymous-auth on by default? The apiserver binary defaults anonymous-auth to true, but kubespray's default was false.
2017-11-06 14:01:10 -06:00
Spencer Smith
19962f6b6a fix indentation for master template (#1906) 2017-10-31 06:43:54 +00:00
Matthew Mosesohn
564de07963 fix indentation for network policy option 2017-10-27 14:56:22 +01:00
Brad Beam
ba0a03a8ba Merge pull request #1880 from mattymo/node_auth_fixes2
Move cluster roles and system namespace to new role
2017-10-26 10:02:24 -05:00
Matthew Mosesohn
b0f04d925a Update network policy setting for Kubernetes 1.8 (#1879)
It is now enabled by default in 1.8 with the api changed
to networking.k8s.io/v1 instead of extensions/v1beta1.
2017-10-26 15:35:26 +01:00
Matthew Mosesohn
ec53b8b66a Move cluster roles and system namespace to new role
This should be done after kubeconfig is set for admin and
before network plugins are up.
2017-10-26 14:36:05 +01:00
Matthew Mosesohn
86fb669fd3 Idempotency fixes (#1838) 2017-10-25 21:19:40 +01:00