RHBA-2018:3789 - Bug Fix Advisory

Topic

Updated director installer packages that resolve various issues are now
available for Red Hat OpenStack Platform 12.0 (Pike) for RHEL 7.

Description

Red Hat OpenStack Platform provides the facilities for building, deploying
and monitoring a private or public infrastructure-as-a-service (IaaS) cloud
running on commonly available physical hardware.

Changes to the instack-undercloud component:

• Previously, undercloud installations with SSL enabled could fail with the error `ERROR: epmd error`. This failure occurred because the VIP that matched the hostname was configured by keepalived after rabbitmq.

With this update, keepalived is configured before rabbitmq, and undercloud installation succeeds. (BZ#1539980)

Changes to the openstack-tripleo-heat-templates component:

• Previously, with shared storage for `/var/lib/nova/instances`, restarting nova_compute on any compute node resulted in owner/group change of the console.log and virtual ephemeral disks of the instance. As a result, instances lost access to virtual ephemeral disks.

With this update, the scripts that modify the ownership of the instance files in `var/lib/nova/instances` no longer cause loss of access to the instance files during restart of nova compute. (BZ#1608913)

• Previously, online data migrations for Nova, Cinder, and Ironic in the overcloud did not run automatically after upgrade to Red Hat OpenStack Platform 12.

With this update, the `openstack overcloud upgrade converge` command triggers online data migrations when upgrading to Red Hat OpenStack Platform 12. (BZ#1627680)

• Previously, users could set RX/TX queue size with `nova::compute::libvirt::rx_queue_size/nova::compute::libvirt::tx_queue_size`. However, there was no dedicated TripleO heat template parameter.

With this update, you can set the RX/TX queue size on a role base:

```
parameter_defaults:
ComputeParameters:
NovaLibvirtRxQueueSize: 1024
NovaLibvirtTxQueueSize: 1024
``` (BZ#1623553)

• Previously, deployments failed with rabbitmq errors when setting the `KernelDisapleIPv6` parameter to `true` due to Erlang Port Mapper Daemon requiring that the loopback interface support IPv6.

To prevent this failure, do not disable IPv6 on the loopback interface when disabling IPv6. (BZ#1631352)

• Previously, the TripleO `CinderNetappBackendName` parameter was not overriding the default value for the cinder NetApp backend correctly.

With this fix, the `CinderNetappBackendName` parameter overrides the default back-end name correctly. (BZ#1599759)

• Previously, the Cinder NetApp back-end driver failed during initialization when the user specified empty string values for some deprecated parameters.

With this update, empty string values for deprecated NetApp parameters are converted to valid NetApp driver configuration automatically and the Cinder NetApp back-end driver initializes successfully. (BZ#1610962)

Changes to the python-os-brick component:

• Previously, the os-brick code that scanned FibreChannel HBA hosts returned an invalid value under certain circumstances. As a result, services including cinder and nova failed.

With this update, the FibreChannel HBA scan code always returns a valid value and cinder and nova no longer fail when scanning FibreChannel HBA hosts. (BZ#1629729)

• Previously, the os-brick code that scanned FibreChannel HBA hosts returned an invalid value under certain circumstances. As a result, services including cinder and nova failed.

With this update, the FibreChannel HBA scan code always returns a valid value and cinder and nova no longer fail when scanning FibreChannel HBA hosts. (BZ#1630842)

• Previously, in certain cases a failure on an individual device flush prevented disconnection incorrectly.

With this update, disconnection fails only when there is a risk of data loss. (BZ#1619792)

• Previously, iSCSI device detection checked for the presence of devices according to the re-scan time. As a result, devices that become available in between scans were not detected.

With this update, iSCSI device detection searches and scans are independent operations at different cadences. (BZ#1629875)

Solution

Before applying this update, ensure all previously released errata relevant
to your system have been applied.

Red Hat OpenStack Platform 12 runs on Red Hat Enterprise Linux 7.5.

The Red Hat OpenStack Platform 12 Release Notes contain the following:

• An explanation of the way in which the provided components interact to

form a working cloud computing environment.

• Technology Previews, Recommended Practices, and Known Issues.
• The channels required for Red Hat OpenStack Platform 12, including which

channels need to be enabled and disabled.

The Release Notes are available at:
https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/12/html/release_notes/

This update is available through 'yum update' on systems registered through
Red Hat Subscription Manager. For more information about Red Hat
Subscription Manager, see:

https://access.redhat.com/documentation/en-US/Red_Hat_Subscription_Management/1/html/RHSM/index.html

Affected Products

• Red Hat OpenStack for IBM Power 12 ppc64le
• Red Hat OpenStack 12 x86_64

Fixes

• BZ - 1460911 - /var/lib/config-data/ irrationally big
• BZ - 1539980 - [OSP12] Undercloud installer fails with ERROR: epmd error
• BZ - 1557327 - Enabling SSL on overcloud endpoints fails
• BZ - 1559676 - ceilometer_agent_centraldocker container doesn't bind /var/log/ceilometer directory with host
• BZ - 1572677 - get-occ-config.sh counts "wrong" cause of missing tonumber screwing up the deployment
• BZ - 1573500 - Keystone still configured for single process
• BZ - 1579693 - [Deployment][RHOSP-12] Lift the restriction on choices for `cpu_model_extra_flags` config attribute
• BZ - 1582409 - [BACKPORT][Deployment][RHOSP-12] Implement CPU weighter filter
• BZ - 1583893 - [RHOSP12] Not able to register baremetal nodes when using pxe_ucs driver
• BZ - 1594764 - [OSP12][Deployment][undercloud] notification_format in nova.conf should be set to unversioned as there is no consumer for versioned
• BZ - 1599759 - CinderNetAppBackendName is not reflected in deployed environment
• BZ - 1608913 - [OSP12] starting nova_compute docker on a new compute node makes guest disks on nfs share read-only
• BZ - 1610962 - [OSP-12] an empty netapp_storage_pools (or empty netapp_volume_list) results in an invalid search pattern
• BZ - 1614451 - [Deployment][OSP12] workaround BZ1613449
• BZ - 1618374 - Backporting PublicVirtualFixedIPs fix in templates
• BZ - 1623553 - [OSP12] RX/TX Queue Size is missing THT parameter
• BZ - 1627680 - [OSP12] Nova, Cinder, Ironic online data migrations not running in overcloud
• BZ - 1627726 - Containerized Cinder logs are not compressed when rotated
• BZ - 1627740 - Allow logrotate configuration to control log retention based on age
• BZ - 1629729 - hp 3par fc volume create from image fails due to Duplicate: Volume (osv-xxx-.xxxxxx) already exists on array
• BZ - 1630842 - fc rescan issue when looking at specific hbas
• BZ - 1643055 - Wrongly rendered namespace in docker images: <class 'jinja2.utils.Namespace'>/centos-binary-ceilometer-notification:current-tripleo"
• BZ - 1643994 - deactivation of interactive login in get-occ-config.sh

CVEs

(none)

References

(none)