This modifies the RabbitMQ Armada charts to wait specifically
for the RabbitMQ StatefulSets to complete prior to proceeding.
This is more robust than the default Pod waiting, as Pod waiting
for SS's has a race condition such that all Pods are likely
Running and Ready prior to any action having been taken, during
and upgrade scenario.
This change leverages recent Armada resource-based wait functionality.
Change-Id: I3867ce6966047d6b877a9bfa872defc0eda9693a
During the initial configuration it's required to
configure a set of valid certificates for ingress.
Make it more explicit so people don't miss this step.
Change-Id: Ie6477f934688467b7d5dfe1cc8191f6acff29a21
The latest version of Armada introduced a separate timeout
value for tests [1], a default value for the timeout is 300,
which is not large enough for Neutron tests.
Also add wait.labels to ensure automatic deletion of test
pods [2].
[1] https://review.openstack.org/#/c/618585/
[2] https://review.openstack.org/#/c/629313/
Change-Id: I86b1f9a8f17c273cf990189b2ffb07d60e3814b9
Make the docs up to date:
* Previous version of the documentation assumes that
partitioning schema is different for SSDs and HDDs,
this is not the case anymore.
* Ceph charts now have automatic partitioning for both
OSDs and journals.
Change-Id: I74bd625522469e2860ada995f4e6a81a566107fa
The version of Armada deployed by Airskiff was recently uplifted. The
gate erroneously uses the latest version of OSH-infra, so there was no
impact; however, users of the 005-make-airship script experience a
Tiller version mismatch due to an older version of K8s. This change
uplifts the OSH-infra pin to prevent errors.
Change-Id: If8f2c27396ce6c8548598492dfaf5e15174131a2
This is to represent specific partitions for journals
instead of whole disk, ceph charts will create journal partitions
and default size is 10GB
Change-Id: I0f032a0d87396fe44f1179b3104ed446cc448d28
This unpins Armada to references that are compatible with helm
version 2.11.0. This is no longer required, as the helm version
used in openstack-helm-infra has been updated to 2.12.1, which
addresses the issue requiring armada needing to be pinned in the
airskiff jobs.
Change-Id: Ibef17b8b0e225b1ec04b6aaebd6f1953dd47b31c
See: https://review.openstack.org/#/c/632794/
Updates to use new Shipyard pod configurations, removes node port use.
Requires Shipyard commit a11e962eef5a5aa8f8fc15c4a324dfa6b2465061 or
newer.
Change-Id: Iff5af45d6b4f60c34d18aad821cf9afa4ea67dda
This change updates the following components:
Kubernetes 1.10.11 -> 1.11.6
CoreDNS 1.1.2 -> 1.1.3 (per k8s 1.11 recommendations)
Etcd 3.2.14 -> 3.2.18 (per k8s 1.11 recommendations)
Tiller 2.10.0 -> 2.12.1 (adds support for k8s 1.11, thread safety fix)
Armada as of 2019-02-4 (adds support for Tiller 2.12.1)
Also fixes tiller upgradability.
Change-Id: I2375b0ae167593815c56470ba6a6c7b529375302
This parameter has been deprecated in Newton and removed [1] in Ocata.
[1] https://review.openstack.org/#/c/385604/
Change-Id: I4e0130fea0375754cebb4a22d2248346f313fe6d
Currently, the Airskiff gate does not recognize failures because the
gate script only verifies that an action lifecycle enters a COMPLETE
state; however, a lifecycle in COMPLETE status is not always an
indication of a successful deployment. This commit updates the
wait_for_shipyard script to check the status of each action step and
results in exit_code 1 if any steps failed.
Change-Id: If5117c4638652b8d5b91fe73ede0f5e19a1e2cc1
Currently, the tests for the OSH rabbitmq charts are disabled; however,
the value that disables the chart is sometimes discarded during document
layering because the `.test` value does not get merged with the parent
document. This commit adds an action to merge the test key and disable
the rabbitmq tests.
Change-Id: Ia23042bd03871cc085f69c232a85375571282af6
Kibana chart is configured with elasticsearch and ldap
credentials, so both kibana password and user are not
used for the configuration.
Change-Id: Ic83087b3bc732d539ac2b837f9c925d0e4ee9849
Previously, Armada was pinned [0] to commit
dd75474d78e69f3634eb94c8ea6730b898e83277 in order to prevent a mismatch
between the Helm API and Tiller. While the generated image tag was
updated, the Armada image deployed by Armada during the bootstrapping
phase was not updated, resulting in a version mismatch. This commit
overrides the tag in order to prevent a mismatch.
[0] https://review.openstack.org/630698
Change-Id: I484c5e52560e5c25c76ad557f3032ae9e158166c
This updates the configuration for the fluent-logging chart to
include inputs for libvirt and qemu, as well as updates the output
configurations to include the appropriate buffer configuration
overrides
Change-Id: Id908ea875b07004202795dac68df7e9727ae9771
* prometheus - change `prometheus` subpath to `monitoring`
* grafana - path to prometheus credentials should be `user`
instead of `admin`
Change-Id: Iee014ce3ff4ee2ebb534003ce3f18ab8bcd63b9d
Zuul