3.2. Red Hat OpenStack Platform 11 Maintenance Releases

This update adds a new user details page to the dashboard.  As a result you can click on the user ID in the Action Log to be taken directly to the user details page.

This update adds a new user details page to the dashboard.  As a result you can click on the user ID in the Action Log to be taken directly to the user details page.

This enhancement provides configuration for OpenStack Load Balancing as a Service (octavia) through the Red Hat OpenStack Platform director.

A high memory consumption was observed, especially in large environments, which often led to out-of-memory issues. The main culprit was neutron-ns-metadata-proxy process, responsible for proxying metadata requests from the VM to Nova.

neutron-ns-metadata-proxy is now replaced by haproxy which has a more lightweight memory footprint.

The OS::Nova::ServerGroup resource now allows you to use the 'soft-affinity' and 'soft-anti-affinity' policies. This is in addition to the 'affinity' and 'anti-affinity' policies.

To use SR-IOV physical function (PF) and virtual functions (VFs) in the same environment, add the 'nm_controlled' and 'hotplug' parameters to the SR-IOV PF configuration in your compute.yaml heat template:

 -type: interface
  name: nic6
  use_dhcp: false
  nm_controlled: true
  hotplug: true

When an OpenStack instance that was using a direct physical function is destroyed, the PCI device is released back to OpenStack and the host system. The root PCI device is then configured to support the number of virtual functions configured during deployment. This process involves the coordination of the host operating system, NetworkManager and OpenStack and may require a short interval of time before the virtual functions are available for use.

Red Hat OpenStack Platform 11 ships with many fewer API workers for each service than in the previous version. Having fewer workers lowers the noise, but also lowers the performance of API response times.

Invalid cache files may cause os-collect-config to report 'ValueError: No JSON object could be decoded' and the service will fail to start. The cache files located in '/var/lib/os-collect-config/' should be valid json files. If they are are of size 0 or contain invalid json, remove the invalid files from '/var/lib/os-collect-config', otherwise they may prevent os-collect-config from starting.

In Highly Available IPv6 deployments, virtual IPs used for RabbitMQ may move between controller hosts during an upgrade. A bug in the creation of these IPv6 IPs causes them to be used as source addresses for RabbitMQ's connections. As a result, RabbitMQ will crash and may be unable to automatically recover its cluster.

To return to normal operation, restart RabbitMQ on the affected controller hosts, as well as any services which depend on RabbitMQ and do not automatically reconnect.

A database upgrade issue hindered the OpenStack Networking (neutron) upgrade. This hindered the upgrade from OpenStack Platform 10 to 11. This fix corrects the neutron database upgrade issue.

This issue affects customers who previously upgraded from OpenStack Platform 9 to 10 and now aim to upgrade to OpenStack Platform 11.

OpenStack Compute (nova) provides both versioned and unversioned notifications in RabbitMQ. However, due to the lack of consumers for versioned notifications, the  versioned notifications queue grows quickly and causes RabbitMQ failures. This can hinder Compute operations such as instance creation and flavor creation. Red Hat is currently implementing fixes for RabbitMQ and director:

https://bugzilla.redhat.com/show_bug.cgi?id=1478274
https://bugzilla.redhat.com/show_bug.cgi?id=1488499

The following article provides a workaround until Red Hat releases patches for this issue:

https://access.redhat.com/solutions/3139721

When using Red Hat Ceph Storage as a back end for both Block Storage (cinder) volumes and backups, any attempt to perform an incremental backup will result in a full backup instead, without any warning.

Previously, during the Red Hat OpenStack Platform 11 deployment, there is a race condition where, on occasion the ceilometer-upgrade runs at the same time as Apache is restarted due to other services being configured in the same step. This resulted in the ceilometer-upgrade to fail as it could not authenticate with the Identity service, as Apache was still not in the active state and aborting the deployment as a failure. 

With this update, as a workaround, when this happens you need to restart the overcloud deploy from where it fails and the deploy should get past this race condition and proceed with the deployment as normal.  As a result, the deployment should be successful instead of failing with an error.

RHEL overcloud images contain tuned version 2.8.
In OVS-DPDK and SR-IOV deployments, tuned install and activation is done through the first-boot mechanism.
  
This install and activation fails, as described in https://bugzilla.redhat.com/show_bug.cgi?id=1488369#c1

You need to reboot the compute node to enforce the tuned profile.

As of this release, Glance API V1 is no longer supported or available.

The store_events option in undercloud.conf has been deprecated and is no longer supported. This option has been removed from the configuration.