This clears out stale YAML structure in preparation of migrating
fresher entrypoints from airshipctl, and authoring of new ones.
Change-Id: I5a006c7e4461b0cce2766b367e5e8494033f8b36
Using MAAS IP for DNS allows installation by not requiring
direct internet connectivity for PXE/bootstrap interface.
NodePort usage for genesis bootstrap helps troubleshooting
during bare-metal deployment. The NodePort is unset once the
full-site (maas-scaled) is deployed.
Change-Id: I6cab0b3cfc541d43d738f3dc144bb29d848ab718
By updating the neutron global document name, it will allow neutron
to be properly replaced at the type and site level as needed.
Change-Id: I2bb734d527aa58fbd0a60252d2630c11f7cd4652
Added airskiff_suse site that deploys the ucp and openstack containers on openSUSE
images. Also renamed the previous airskiff site repo to airskiff-ubuntu. Added new
gates for airskiff-suse lint and deployment.
Change-Id: I5672e8db556ef02c3d74a598d2772fbebccf6f77
Now Shipyard [1] requires `statefulset` to be configured
for airflow endpoints.
This patch fixes both airskiff gates and other sites.
[1] https://review.opendev.org/#/c/671575/
Change-Id: I29edeb27763059b1dd5f99db4740e0913f0dc56d
Create a site type called skiff and point airskiff site to be based on
the new type. This will help to reduce the duplication when creating
airskiff-suse site in subsequent commit.
Change-Id: Ie9cd6d0e572a4556ee6ff5c3d37f20349c2d7de3
This site is created in order to utilize the multinode development
environment alongside the global and type manifests in Treasure map.
To accomplish this, the new seaworthy-virt site is a copy of the
airship-seaworthy site but with as many overrides and removed
pieces as necessary.
Change-Id: I6d19e1cf019c5d03f42343ab3c72971172879e4d
- uplift ceph charts
- This is to enable rack level failure domains as we are running on
host based failure domains
- updating the site to have one override which will create custom domain
names since current logic uses hostnames to get domain names.
- this lab running on single rack so using site override to create custom failure
domain names.
Change-Id: I2b92055e77dac6eba4fe59694c5291046149a6c4
Warning: Nameserver limits were exceeded, some nameservers have been omitted,
the applied nameserver line is: 10.96.0.10 8.8.8.8 8.8.4.4
Removing 3rd DNS server entry fixes the issue.
Change-Id: I14fae0d884223050eb3106b4a9cf1adde01def8f
This change integrates changes related to the new Patroni-enabled
high availability included in the postgresql helm chart.
1. Updates the version of the postgresql chart
2. Adds users/passphrases for the postgresql replication and exporter users
3. Sets the default number of replicas to 3 (except for skiff & sloop)
4. Adds documentation/scripting for upgrading the chart
Co-Authored-By: Evgeniy L <eli@mirantis.com>
Change-Id: I51e63fb2444610cfb144bc587a6a179fa4830809
The intention for the patch is to increase the stability of
OpenStack installation:
1. Use a single RabbitMQ cluster for all OpenStack services.
2. Use 2 replicas for RabbitMQ clusters instead of 3; RabbitMQ
does not implement a quorum mechanism which allows us to have
even number of nodes, this should provide better performance
by reducing the number of instances where the data should be
replicated. This should also reduce the odds of getting
RabbitMQ partitioning.
Co-authored-by: Drew Walters <andrew.walters@att.com>
Change-Id: Id981e2650efc24a9e67cf87a0abd414a96f40088
After enablement of etcd encryption in https://review.opendev.org/628290 for
promenade, bootstrap-armada pod should be run with
'--experimental-encryption-provider-config' service parameter to avoid issue:
'ERROR: No API token found for service account "airship-ucp-ceph-osd-test"'
Change-Id: Ib9bf1fa7333874b2d88db84019b26a2691a7d18a
fluent-logging chart was splitter into fluentbit and fluentd charts
in the following PS: https://review.opendev.org/642146
Change-Id: Ic72f9fb6293b99395a0acf48064b9c04e312d365
BaremetalNode/v1.yaml: volume_group is
defined twice in same section in lines 58 and 66.
HostProfile/v1.yaml: volume_group is
defined twice in same section in lines 49 and 57.
SoftwareVersions/v1.yaml: type is defined twice
for dep_check
config/service_accounts.yaml: user_domain_name
defined twice in same section in lines 109 and 114.
config/service_accounts.yaml: user_domain_name
defined twice in same section in lines 111 and 116.
Change-Id: I1597c0afc6b3a9674c0c74692f6575d797bd714d
By specifying exact ubuntu release as a tag for `updater.py` script, it prevents
downloading ubuntu with a higher release, for example, ubuntu bionic.
Changing bionic to xenial for armada chart.
Change-Id: I1864c3f8db2d90ab63d7d923de8f2ac0ecd20e31
The default image tags recently changed for the Deckhand image [0]. This
change moves to use the default tag.
[0] https://review.opendev.org/638301
Change-Id: Ia7198ec579bb68e6fd64e00041ea6db366adf341
Add explicit glance and heat overrides, this fixes problems
when two manifests with identical names are present,
one manifest comes from the type level, another one comes
from the site level, this leads to random manifest being
picked up during service deployment by Armada.
Change-Id: I87f6cce910dcf1f87370ba231077451990b245ee
Note: all scripts in tools/deployment/aiab directory have been
moved into this repository as is (except a few changes to make
them work), they require a heavy refactoring that will be done
in separate patch-sets.
* Add a virtual single node manifests that are based on a sloop type.
* Use NFS provisioner instead of Ceph.
* Update tools/openstack to be non-seaworthy specific, use a default
region name and auth url.
* Make type/sloop/config/common-software-config.yaml to be site specific,
to allow to configure custom region_name.
* Remove max-pods-per-cpu parameter for kubelet, treasuremap needs
to support a diverse set of environments, without constraints
on the number of available cores. Max pods configuration parameter
is still present and helps to mitigate problems when kubernetes
unexpectedly starts a large number of pods.
Change-Id: I379a50d810b91b989f039dbb7c691f5ceec0cc67
Since support for additional operating systems has been added to some
components, the image tag "latest" is no longer used by their Makefiles.
This commit changes the tags for the aforementioned components to use
the new default tag and updates the job name to reflect the use of
Ubuntu images.
Change-Id: Ia1adf156c4b3798b2400f9c821ac2ba685557591
Currently, Airskiff deploys a custom manifest defined in
tools/deployment/airskiff/manifests/airship.yaml to boostrap Airship
components. This exempts Airskiff from the daily version uplift process
and causes a divergence from the global UCP charts. This commit updates
the Airskiff bootstrap process to use the existing UCP charts at the
global layer.
Change-Id: I68dc405fda3cb1dbf5865104b657bdd75812e9bc
Override default "metadata_proxy_shared_secret" parameter.
This secret is used by Neutron to sign instance-id headers
to prevent spoofing when proxying metadata requests.
Change-Id: I771d7f818a18b82d55bf781d71fc95114ac7e78c
Sloop type/site is a minimalistic approach to Airship with reduced
requirements towards hardware and external dependencies while keeping
all the functional features.
Major differences compared to reference site airship-seaworthy
- Two bare-metal server setup with 1 control, and 1 compute.
Most components are scaled to a single replica and doesn't carry
any HA as there is only a single control plane host.
- No requirements for DNS/certificates.
HTTP and internal cluster DNS is used.
- Ceph set to use the single (root) disk. This generally
provides minimalistic no-touch ceph deployment.
No replication of ceph data (single copy).
- Simplified networking (no bonding).
Two network interfaces are used by default (flat PXE, and DATA network
with VLANs for OAM, Calico, Storage, and OpenStack Overlay)
- Generic hostnames used (airsloop-control-1, airsloop-compute-1) that
simplifies generation of k8s certificates
- Usage of standard Ubuntu 16.04 GA kernel (as oppose to HWE)
Change-Id: I4944fcae7d29ed8799d810c93efb0120b6b3a105
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
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
Updates to use new Shipyard pod configurations, removes node port use.
Requires Shipyard commit a11e962eef5a5aa8f8fc15c4a324dfa6b2465061 or
newer.
Change-Id: Iff5af45d6b4f60c34d18aad821cf9afa4ea67dda