GitLab

Sync arcus/train-staging changes into arcus/train-production

See merge request !131

Ensure FQDN is configured correctly

See merge request !129

Tweak inventory around faulted and fixed nodes

See merge request !130

Remove some hard-coded node fixes around the FQDN issue and make staging env DNS config consistent with production

This is using the following patch:

https://review.opendev.org/c/openstack/ansible-collection-kolla/+/852031/3/doc/source/roles/baremetal.rst#20

Back ported to our kolla-ansible fork. You will need to do a control host bootstrap/upgrade to
update your kolla-ansible.

Sync changes from arcus/train staging

See merge request !128

10x IRIS hosts for CCFE-DAKOTA do no need CPU isolation configs

See merge request !127

Sync arcus/train-staging changes into arcus/train-production

See merge request !126

Add 10x CCLake nodes for new IRIS 2022/23 capacity

See merge request !125

Sync arcus/train-staging changes into arcus/train-production

See merge request !124

Reimage staging node svn2-ar17 and configure as a nova-compute hypervisor

See merge request !123

# Conflicts:
#   etc/kayobe/ansible/nova-compute-drain.yml

Update drain script to do cold migration

See merge request !122

Bump neutron containers + Fix etcd backend_url

See merge request !121

(cherry picked from commit 953a299e9fb0c8c5c867f22b6d400b04e4f03062)

Fix Cinder etcd coordination URL

See merge request !120

Bump neutron containers

See merge request !119

Disable svn2-ar17-u21 in inventory temporarily while the node is re-imaged

Cinder etcd coordination url should point to the controler-specific Internal API network address, and not that of controller[0] as in upstream Kolla-Ansible

- https://github.com/stackhpc/kolla/commit/cd4e843638f92c64ad8db8caa5dccf7f5f089f51
- https://gitlab.com/scientific-openstack/infrastructure/kayobe-config/-/commit/265137dee285935e718062167e14db0cb458e681

Build logs:

- centos 7: https://gitlab.com/scientific-openstack/infrastructure/kayobe-config/-/jobs/2727956842/artifacts/browse/artifacts/
- centos 8: https://gitlab.com/scientific-openstack/infrastructure/kayobe-config/-/jobs/2721852494/artifacts/browse/artifacts/

Apply Nova and Neutron tunings as attempts to work around etcd locking issues

See merge request !116

Bump neutron server tag

See merge request !117

Revert production deploy load tunings

See merge request !118

This fixes an issue with batching, see:

https://gitlab.com/scientific-openstack/infrastructure/kayobe-config/-/commit/d3b1a69746a2d6c7ecf58e31c1731a00305cade3

Sync from scientific-openstack/train

See merge request !102

# Conflicts:
#   etc/kayobe/environments/builder-centos7/centos7.yml