20.10

Summary

The 20.10 OpenStack Charms project release includes updates for the charms described on the charm guide page. As of this release, the project consists of 60 supported (stable) charms.

For the list of bugs resolved in this release refer to the 20.10 milestone in Launchpad.

For scheduling information of past and future releases see the Release schedule.

General charm information is published in the OpenStack Charm Guide (this guide) which ultimately supersedes Release Notes contents.

Important

Always upgrade to the latest stable charms before making any major changes to your cloud and before filing bug reports. Refer to section Upgrading charms below for details.

New charm features

With each new feature, there is a corresponding example bundle in the form of a test bundle, and/or a section in the OpenStack Charms Deployment Guide, that details its usage. Test bundles are located in the src/tests/bundles directory of the relevant charm repository.

OpenStack Victoria

The charms now support OpenStack Victoria on Ubuntu 20.10 (natively) and Ubuntu 20.04 LTS (via UCA).

Gnocchi and external root CA for S3

The gnocchi charm now supports an additional configuration option: trusted-external-ca-cert. This option installs and configures an external root CA in the gnocchi units. This is useful when an encrypted S3 (storage backend) endpoint uses certificates that are not managed by Vault.

Arista - Supported releases

The neutron-api-plugin-arista charm now supports all OpenStack releases starting with Queens. See bug LP #1890628 for more details.

Neutron Gateway and additional data on ports and bridges

The neutron-gateway charm now marks the openvswitch ports and bridges it creates as managed by the charm. It does so by setting an external-id on them. See the Open vSwitch Integration Guide for Centralized Control for details.

Even the ports and bridges created before upgrading to this release will be marked. Marking our ports and bridges will allow us to implement some cleanup mechanism without risking removing too much.

See bug LP #1809190 for details.

cinder charm: Allow specifying the default_volume type

A new feature has been added to allow setting the default volume type, via the new configuration parameter default-volume-type, for new volumes if one is not specified at creation time. This is to support scenarios where multiple storage backends are to be used with cinder. Please see related bug LP #1884548 for more details and consult the cinder charm.

Action based migration from Neutron ML2+OVS to ML2+OVN

New actions have been introduced to enable migration of a Neutron ML2+OVS based deployment to ML2+OVN.

Please refer to the Migration to OVN page for more information.

Ceph BlueStore compression support

Charms that allow for the configuration of Ceph storage can now optionally enable Ceph BlueStore compression for the pools they manage. These are:

  • ceph-fs

  • ceph-radosgw

  • cinder-ceph

  • glance

  • gnocchi

  • nova-compute

Configuration options have also been added to the ceph-osd charm which allow overriding compression settings for the deployment as a whole.

Please refer to individual charm documentation for a description of the configuration options.

New charms

ceph-iscsi

The ceph-iscsi charm is now a supported charm. This charm implements the Ceph iSCSI gateway, which provides iSCSI targets backed by a Ceph cluster.

Preview charm features

n/a

Deprecation notices

ceph-radosgw charm: Deprecate the pool-prefix config option

The pool-prefix configuration option for the ceph-radosgw charm is deprecated and will be removed. See bug LP #1856106 for a full discussion.

Octavia: Spares pool support is deprecated

Spares pool support is deprecated upstream, pending removal in the X release. The charm will retain support for the feature for OpenStack releases up until the X release. Please refer to Octavia deprecation notes for Victoria for more information.

Removed features

Neutron Gateway and network bridges

The neutron-gateway charm no longer supports adding a Linux bridge to an openvswitch bridge. This ability was dependent upon the veth peer links feature of NetworkManager (a.k.a. ifupdown). Yet starting with Ubuntu 18.04 LTS (Bionic) ifupdown has been replaced by Netplan, which has a feature gap in this area (see bug LP #1876730).

Using ifupdown on Bionic also causes issues with LXD containers (see bug LP #1877594). The latter issue has details on migrating away from veth peer links.

Removed charms

n/a

Known issues

Barbican DB migration

With Focal Ussuri, running command barbican-manage db upgrade against a barbican application that is backed by a MySQL InnoDB Cluster will lead to a failure (see bug LP #1899104). This was discovered while resolving bug LP #1827690.

Both the charm bug LP #1827690 and the package bug LP #1899104 are known issues that will be addressed shortly after the 20.10 release.

The package bug only affects Focal Ussuri and is not present in Victoria, nor is it present when using (Bionic) Percona Cluster as the back-end DB.

Designate and Vault at Ocata and earlier

The designate charm for OpenStack Ocata (and earlier) does not yet support SSL via Vault and the certificates relation. See bug LP #1839019. The charm works as intended in this scenario starting with OpenStack Pike.

TrilioVault Data Mover charm upgrade

For deployments using prior versions of the trilio-data-mover charm (as provided by Trilio) the relation between the trilio-data-mover charm and rabbitmq-server must be removed and re-added to ensure that specific access for the data-mover service is provided for RabbitMQ.

juju remove-relation trilio-data-mover rabbitmq-server
juju add-relation trilio-data-mover rabbitmq-server

TrilioVault File Recovery Manager

Mounting snapshots using the File Recovery Manager appliance fails due to permissions errors encountered during the libvirt/qemu snapshot mount process on compute nodes. See bug LP #1888389 for details.

Octavia and neutron-openvswitch in LXD

The octavia charm requires a neutron-openvswitch subordinate which means that if it runs in a container, the openvswitch kernel module must be loaded before the container starts. Module loading is done by LXD based on the profile applied by Juju and taken from the neutron-openvswitch charm. However, due to a combination of bugs (LP #1876849 in Juju and LP #1906280 in the ovn-chassis/neutron-openvswitch charms) there is no guarantee that the profile will be applied before neutron-openvswitch (or ovn-chassis) execution starts in a container.

The issue is more likely to happen on disaggregated deployments where octavia units run in LXD containers on machines that do not have any units of neutron-openvswitch running on bare metal.

In order to work around the error an operator needs to make sure the openswitch module is loaded on the host and then restart the openvswitch-switch.service service inside the LXD container where the respective neutron-openvswitch unit is present. After that the unit error can be resolved.

OpenStack os-brick, Ceph Octopus, and Focal

The Ceph RBD Mirror and Cinder Backup Swift Proxy charms do not work with Ceph Octopus due to an issue with the upstream OpenStack os-brick library (see bug LP #1865754). As Octopus is the default Ceph version on Ubuntu 20.04 LTS (Focal) these charms cannot be used on Focal until the issue is resolved. Here are the resulting charm-specific behaviours:

  • ceph-rbd-mirror charm: The charm will enter a blocked state after configuring pool mirroring (see bug LP #1879749).

  • cinder-backup-swift-proxy charm: If a backup volume operation is performed the resulting volume will be in error (see bug LP #1890821).

Series upgrade - percona-cluster and vault charms

percona-cluster

During a series upgrade from Xenial (16.04) to Bionic (18.04) the percona-cluster charm may fail during the post-series-upgrade hook. This appears to be because the percona-cluster charm may erroneously delete the file /var/lib/percona-xtradb-cluster/seeded (see bug LP #1868326). If this occurs, then executing the following commands on the failed unit will recover the hook and allow it to complete the series upgrade:

juju run percona-cluster/N 'echo "done" > /var/lib/percona-xtradb-cluster/seeded'
juju resolved percona-cluster/N

This may be required for each percona-cluster unit.

vault

If a series upgrade is attempted while Vault is sealed then manual intervention will be required (see bugs LP #1886083 and LP #1890106). The vault leader unit (which will be in error) will need to be unsealed and the hook error resolved. The vault charm documentation includes unsealing instructions, and the hook error can be resolved with:

juju resolved vault/N

Upgrading charms

Always use the latest stable charm revision before proceeding with topological changes, application migrations, workload upgrades, series upgrades, or bug report filing.

Please ensure that the keystone charm is upgraded first.

To upgrade an existing deployment to the latest charm version simply use the upgrade-charm command. For example:

juju upgrade-charm keystone

Charm upgrades and OpenStack upgrades are functionally different. Charm upgrades ensure that the deployment has the latest charm revision, containing the latest charm fixes and features, whereas OpenStack upgrades influence the software package versions of OpenStack itself.

A charm upgrade does not trigger an OpenStack upgrade. An OpenStack upgrade is a separate process. However, an OpenStack upgrade does require the latest charm revision. Please refer to OpenStack upgrades in the OpenStack Charms Deployment Guide for more details.