* DIB element is updated to the latest k8s version - 1.3.0
* K8s application is fixed according to the new k8s version:
- replaced short '-t' flag on the long one - '--template'
Change-Id: I37d73a907c55c0e3aca6a2e786b05a9e99a215ac
This patch adds actions to KubernetesPod class which allow to
recreate a pod (delete it and create new one) and to restart docker
containers on minion nodes.
Change-Id: Ia03b27d72662700fb302960adba9dae20203b597
Closes-bug: #1466359
During scaling gateways down, reporter was trying to get ip of the
instance which was already released and that caused index error.
This patch saves this ip for report before releasing resourses.
Change-Id: I9310a423b1ad8c488c1cc7400844ae5d8767879c
Closes-bug: #1590728
* Increases format of packages
* Adds missed licence headers
* Other minor refactoring
Change-Id: I580bea2eb2f7e4616cd12245367ada8d43c5c4ff
Targets: bp refactor-apps-scripts
Allow access to Google Container Registry using json key as
a password for docker login. The key is scoped to a specific
Google Cloud Platform Console project and its resources.
Add new field for key.
Partially implements: blueprint k8s-google-registry-support
Change-Id: Ica71d1a48ac7fb75a71497fcbe318d1b9e4e8192
Allow access to private docker registry using docker argument
'--insecure-registry'.
Add new field for docker private registry host ip or domain name.
Partially implements: blueprint k8s-docker-enterprise-hub-support
Change-Id: I725c0b6a2bdbfe6eb9e627266411286fe44bfae3
1. There was scaleGatewayUp but no Down action
2. Gracefully remove node from etcd upon down-scale of nodes/gateways
3. Log to user which gateways were created/deleted
4. Deployment either hanged or worked incorrectly when downscaling and then
upscaling something again
Also sets minimum gateway node count to 1 until gateway-less mode be fully
functional again
Change-Id: Ie3255f70b8a3fdfadab119a579c00433938495c8
Closes-Bug: #1509074
Some images are updated and may not supported in the future.
Murano admins should build images by there own
from fresh elements, provided in the application folder
Change-Id: I683bcac64e46e4c93a1b958e5b79452874c44288
(cherry picked from commit bd51f268525a55c7800a32a3e3d3fd5e9ffc4fb0)
When Environment get deleted both Docker/k8s host and
Docker applications get deleted. So it is possible that host
will try to delete container from an Instance that was already
destroyed. Attempt to send commands to such instance
will hang until timeout because there is no more agent
exist on the other side.
Change-Id: Ia6e8064ae50c24ce595ed08e446ed45b96b4a330
Closes-Bug: #1504623
Depends-On: I298f3f079a153ef87ed7213b7f1432c9e5e4b4b6
* Kubernetes element was updated to base on v1.0.6
* Docker now has its own repo and has different package name.
Also it is now installed via downloadable script.
TODO: Need somehow to fix docker version to 1.8.2
Change-Id: I230e9705cb850e9fb3772a67284561907d4be615
Implements: blueprint update-k8-to-v1-dib-elements
Co-Authored-By: Alexey Deryugin <aderyugin@mirantis.com>
etcd services on Kubernetes nodes should not start concurrently
or even close to each other, otherwise a split-brain may occur.
This patch increases the delay between service starts from 10 to
30 seconds.
Change-Id: I76a6de47e1fff911373cee1909580be73bf8fb02
Partial-bug: #1494999
The script which adds new nodes to etcd cluster now does up to 30 attempts
and fails (returns a non-zero exit code) if none are successful.
A minor performance optimization has been done as well, so no unneeded
waits happen when the command is successful and no temporary files are used.
Closes-Bug: #1494997
Change-Id: Ic0552f388518c119925da902b64797bbf96f979f
Availability zone was present for master node but not for
minion and gateway nodes which caused error when AZ was
not "nova"
Change-Id: I5b223f20223c2d4d1f2d797f94cd366e206773da
Closes-Bug: #1490575
Due to the fact that yaql 0.2 allows to compare the number with a
string there are no errors in this check and method works
incorrectly allowing to application to use port 22
Change-Id: I4e6be454b9ee3d6f5f594856610c201e741f310d
Closes-Bug: #1481748
Replace call of undefined function 'lowercase' with standart bash string conversion to lowercase
Change-Id: I7fa4a85058e81afbd72d2f245c66995afc5ce26a
Closes-Bug: #1465282
Application deletion were supported on DSH/Pod but not exposed
in DockerApplication. Once it became exposed number of bugs in
KubernetesPod and DSH became visible.
Change-Id: I33650936a73994668f72348a909b7a51bae2b98c
Closes-Bug: #1446582
(cherry picked from commit a74293e2b4ed5695737066d9154d51a4170b8a80)
UpdatePod execution plan was using field from Pod
definition that was present in v1beta1 API but is no
longer used in v1beta3
Change-Id: Ib57a0e5cc7104d5e1fa39b07644eab182a7fdf45
Closes-Bug: #1447594
(cherry picked from commit a680bf5899e006ff1e3ecf1ee3e3c7cce4a9bebb)
KubernetesCluster used to report 2 CPU cores and 4GB RAM
for each Node no matter what flavor was used for the Node.
This fix adds additional step to ask VM for its real capabilities
and provide those values during Node creation
Closes-Bug: #1447386
Change-Id: I786de0b5fcf5cf094397c8c88b2f03a88326b884
(cherry picked from commit dee685beaba7270b0258c8a5be2d028ecab6d2f8)
Execution plan to patch existing k8s services to include node IPs
when number of nodes changed and we are not using gateways
or we was using gateways and now we don't
These 2 files were supposed to be in
https://review.openstack.org/#/c/175506/
but I forgot to commit them
Also fixes typo that is already fixed in stable/kilo
Change-Id: I6af644511bf25ae7ae31c88914d33b8c518a4eec
