In MariaDB we adjust the dependecies of what we install there as well
adding only what is appropriate for the install.
In Ansible we adjust some templates to all work around differences
between the different Linux families.
Change-Id: Ibc26e2f4d4a732630632d3ed27fb595b6fe019d2
Partially-Implements: blueprint install-from-ubuntu
Fixes issue where keepalived will not restart after container failure.
Removes keepalived.pid file if one exists on container startup.
Change-Id: I6ec3297a5e49a4d5255323aa7f75dfdc10d0c5ec
Closes-Bug: #1485079
Currently we write configs to /opt/kolla/config on the destination
node. This does not mesh with LSB so we adjust it to /etc/kolla
Change-Id: I4b6683b3b707ebb25b064ba5162533ea5a35d977
Partially-Implements: blueprint update-configs
Docker cache takes into account things like filesize, hash of file,
and even timestamp and access times when determining whether to use
the cache or not. This modifies the build script to adjust all the
timestamps to 0 epoch. Since we dont care about timestamps this allows
us to make far better use of the cache, including when downloading and
creating tarballs for source builds.
Change-Id: Id4287cf1fcaa2de63feaab9e6f4ebbd4defdb768
Partially-Implements: blueprint build-script
In rabbitmq we removed the requirement for hostname since that is
already installed.
Change-Id: Ia2355e332d53763db630d19663c2ddc8bf3aa425
Paritially-Implements: blueprint install-from-ubuntu
In the base we no add the sources.list with the appropriate sources
from around the interwebs. We also pull in the appropriate signing
keys for those packages. We adjust the dependencies as well.
Change-Id: Ic7e084c0c57287945d0e7c2526f050afbbdec454
Partially-Implements: blueprint install-from-ubuntu
The register operation was registering user_database which is a global
variable in group_vars/all.yml. This results in glance not being deployable
because user_database is overwrriten at keystone task registration time.
The result of this is the playbooks fail to deploy.
Change-Id: Ic69725fb09aa0a368ec5b0d8f47bfc768c29c70c
Closes-Bug: #1486876
If ip_vs is not loaded keepalived will enter a restart loop. To fix that
load ip_vs module every time we start the container.
Change-Id: I6b39d8654cc6a6da7a9162e5c3efab073746b246
Closes-Bug: #1486276
* Enable some extra repos for oraclelinux in the base template.
* Disable magnum binary builds for oraclelinux due to unavailable
python-docker-py dependency
Change-Id: Iaf8826dc0f9ff93a40e5ad3d1c0fd1fb6ef2e048
Closes-Bug: 1486636
We now depend on jinja2 and docker-py for building images.
We no longer depend on docker-compose for any part of our
runtime so removing it as a requirement. If we keep compose
in the tree, it will be for reference only, not for runtime.
Change-Id: Ie663a0f183e74c14f01e9cb4fdc8763355c6415a
Partially-Implements: blueprint dockerfile-template
Libvirt uses libvirtd.conf not libvirt.conf. Currently,
the config file get copied to '/etc/libvirt/libvirt.conf' and
libvirt creates its own libvirtd.conf and ignores the
configured libvirt.
Change-Id: I7a55b92c33ade72b7ac8ba028c67313479ad8343
Co-Authored-By: Ian Main <imain@redhat.com>
Closes-Bug: #1486100
The previous method of using the API image for a data container
worked, but resulted in more data being bindmounted and copied
into the system. This patch resolves that by essentially using
a distro-type as a FROM in a Dockerfile to generate
distro-type-data. Further all of the data containers used
throughout ansible are modified to use this new data container.
Change-Id: I8846573d4f2d4b98d4f46c770bfefc6d4c5cd0b5
Partially-Implements: blueprint one-data-container
