RHBA-2019:2963 - Bug Fix Advisory

Topic

A new release of Red Hat Hyperconverged Infrastructure (RHHI) for Virtualization that fixes several bugs and adds various enhancements is now available.

Description

Red Hat Hyperconverged Infrastructure (RHHI) for Virtualization integrates
Red Hat Virtualization and Red Hat Gluster Storage, allowing you to
consolidate deployment and management of compute and software-defined
storage resources.

The Red Hat Virtualization Host (RHV-H) package has been upgraded to
upstream version 4.3-20190925.7. This includes Red Hat Gluster Storage
(glusterfs) packages that are upgraded to upstream version 3.12.2-47.5.
These packages provide a number of bug fixes and enhancements.

This update fixes the following bugs:

• The gluster service was unnecessarily restarted during host

activation, causing volumes to lose quorum and stop working.
This issue is prevented by checking the state of the gluster
service during host activation and starting only when required.
(BZ#1751717)

• When more nodes were added to a hyperconverged cluster using

'Expand cluster' in the Web Console, the new nodes were not added
to the trusted storage pool, and could not be immediately used for
storage. New nodes are now added to the trusted storage pool when
'Expand cluster' is used in the Web Console. (BZ#1693144)

• Gluster hook scripts are now verified using SHA256 for checksum,

as the previous MD5 checksum was not appropriate for FIPS enabled
systems. (BZ#1723676)

• All files are now correctly removed during cleanup, so the

'Use existing deployment' option is no longer visible after cleanup.
(BZ#1640868)

• If pending heals exist or gluster quorum cannot be maintained

when you attempt to upgrade an online host, the host upgrade fails
so that data or quorum are not lost. Log messages are now more
informative when hosts cannot be moved to maintenance, for example:
Cannot switch the following Host(s) to Maintenance mode: server1.example.com.
Gluster quorum will be lost for the following Volumes: myvolume.
(BZ#1649502)

• Hostname validation now includes hostnames specified in the

/etc/hosts file. (BZ#1692671)

• The cleanup playbook now correctly detaches gluster peers and

completes successfully. (BZ#1692782)

• Hyperconverged hosts are added to the Red Hat Virtualization

default cluster using only the front-end FQDN, preventing failures
that arose from also using the back-end FQDN. (BZ#1715952)

• LVM cache configuration did not work with Ansible 2.8 because of

changes to expected values for 'pvs' in the configuration file.
This has been updated so that the correct syntax is used. (BZ#1734778)

This update also provides the following enhancements:

• Thinly provisioned bricks are now supported on top of Virtual Disk

Optimization (VDO) devices. (BZ#1693653, BZ#1693653)

• Multipath devices are now supported for use with

RHHI for Virtualization. (BZ#1720334)

• Volume expansion through adding a new brick to an existing volume

is now supported in the Web Console. (BZ#1720918)

• Failed storage deployments can now be cleaned up from the

Web Console. (BZ#1721370)

• LVM cache can now be attached to a specific thin pool, instead of

only the first thin pool. (BZ#1721451)

• Previously, when deployment checks failed, deployment failed after

some time instead of failing immediately. The checking method has
changed so that any unmet prerequisites stop the deployment process
immediately. (BZ#1686357)

• More verbose (-vv) logging can now be enabled in the Web Console

during deployment. Check the 'Enable debug logging' checkbox to use
this feature. (BZ#1734732)

Red Hat recommends that all users of RHHI for Virtualization upgrade
to these updated packages, which fix these issues and provide these
enhancements.

Solution

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

For details on how to apply this update, refer to:

https://access.redhat.com/articles/11258

Affected Products

• Red Hat Gluster Storage Server for On-premise 3 for RHEL 7 x86_64

Fixes

• BZ - 1640868 - [gluster-ansible] Include ansibleStatus file removal with the gluster configuration cleanup
• BZ - 1649502 - No proper events or error message notified, when the host upgrade fails
• BZ - 1674598 - Remove the validation on acceptable latency during gluster deployment
• BZ - 1679863 - UI exception seen in RHV-M when there is any failure/error
• BZ - 1686357 - [gluster-ansible] Fail at the very beginning, when pre-requisites for RHHI-V deployment is not met
• BZ - 1690797 - Disk count and stripe size are missing in the deployment wizard, when changing disktype from JBOD to RAID
• BZ - 1692652 - Playbooks to create gluster bricks are missing
• BZ - 1692671 - RHHI Gluster deployment fails for static hostnames
• BZ - 1692782 - Peers are not disconnecting after gluster cleanup using playbook.
• BZ - 1692791 - [Day 2] While creating new volume or expanding the volume using day2, option to add brick should be removed
• BZ - 1693144 - [Day 2] Peers are not probed during expand cluster.
• BZ - 1693653 - Enable LV thinpool on VDO devices
• BZ - 1714781 - Partitions are allowed only for PoC mode for deployment.
• BZ - 1715458 - Changing the name from FQDN to additional hosts in cockpit
• BZ - 1715952 - Single node RHHI-V deployment, results in the host added twice to the cluster, one with backend and other with frontend FQDN
• BZ - 1720334 - Disable blacklisting all devices during gluster deployment
• BZ - 1720918 - Cockpit Day 2 - RFE - Support for expanding volume/creating bricks
• BZ - 1721370 - Provide a way to cleanup gluster deployment from cockpit
• BZ - 1721451 - [RFE] Provide option for user to attach LVM Cache to specific thinpool
• BZ - 1721475 - Improvements on Gluster Deployment in RHHI
• BZ - 1723676 - Change md5 checksum used in GlusterHooks calls
• BZ - 1723678 - Typo and exception due to non-iterable object on gluster fencing testing
• BZ - 1734376 - Mount the bricks with XFS UUID instead of device names
• BZ - 1734386 - Update the VDO systemd service file to support thinp bricks
• BZ - 1734717 - Update LV cache related variables in generated vars file
• BZ - 1734732 - [RFE] Add Log verbosity flag to Cockpit installer
• BZ - 1734778 - LVM cache feature not possible with ansible-2.8
• BZ - 1745483 - Deployment wizard has overlapped buttons, for the scenario of reusing existing configuration
• BZ - 1745498 - VDO with LVM thinpool needs correct entry in inventory file in cockpit based deployment.
• BZ - 1745560 - Remove the default arbiter check for vmstore volume
• BZ - 1746394 - Additional hosts not getting added in RHV-M
• BZ - 1746451 - lvcache should be the same as input from user
• BZ - 1746724 - Day 2 [Expand Volume]: Logical size is incorrect while providing the LV size
• BZ - 1751717 - Host activation causes RHHI nodes to lose the quorum
• BZ - 1751722 - Gluster fuse mount crashed during truncate
• BZ - 1753096 - Rebase RHV-H on glusterfs-3.12.2-47.5.el7rhgs

CVEs

(none)

References

(none)