2.2. Known Issues

These known issues exist in Red Hat Enterprise Virtualization 3.1 at this time:

Administration Portal

BZ#868964
Due to the performance constraints imposed by the JavaScript engine included in Internet Explorer 8 it is not a supported client for the new Administration Portal user interface. Later versions of Internet Explorer remain supported.

Application Programming Interfaces

BZ#830517
The new Python SDK maps actions and properties directly to the REST API with the exception of import actions. As import is a reserved keyword in Python these actions have instead been renamed import_resource, where resource is replaced by the resource being operated on, for example vm or template.
BZ#830913
A new syntax for performing event search using the REST API is now supported. The event search URI is now of the form /api/events;from=<event_id>. The existing search syntax remains supported but is considered deprecated and will be removed in a future release.

Backend

BZ#874443
In previous releases when the disk image of a virtual machine was being modified or used by a task the entire virtual machine was locked. This lock is now more specific and applies to the disk. This means that users are still able to perform additional tasks on the virtual machine, such as adding or removing disks, concurrently. 
				
As a result of this change of behavior the virtual machine status may indicate that a virtual machine is "Up" when the attached disk is in fact being used by another task. Users will not be aware of this lock until such time as they attempt to perform an action that requires a lock on the disk.
BZ#856273
On large scale deployments, of 200 hosts and above, you may need to increase the maximum allowed connection on the database server from the default value of 150 to 75% of the expected number of hosts. By default this value is found in the "/var/lib/pgsql/data/postgresql.conf" file on the database server.

Directory Services

BZ#814445
It is not possible to install Red Hat Enterprise Virtualization Manager (rhevm) and IPA (ipa-server) on the same system. IPA is incompatible with the mod_ssl package, which is required by Red Hat Enterprise Virtualization Manager.
BZ#739431
Red Hat Enterprise Virtualization Manager does not support the use of user accounts which are marked as expired in the directory server for the domain. This applies to both Active Directory and IPA domains.
				
The password for the account must be changed in the directory server, re-enabling the account,  before it is used with Red Hat Enterprise Virtualization Manager.
BZ#828877
Use of the IBM Tivoli Directory Server (ITDS) provider option of the "rhevm-manage-domains" command is not supported in this release.

Migration

BZ#731100
Migration of a virtual machine between hosts that do not have their clocks synchronized may lead to the virtual machine being unresponsive for a period of time following migration. It is important to ensure that all hosts in your virtualization environment have their clocks synchronized using NTP.
BZ#813279
If the VDSM process is restarted while moving a virtual machine between storage domains the task is currently only partially rolled back. This leaves the virtual machine in an illegal state, unable to be started or moved again. If this occurs contact Red Hat Support.
BZ#844382
Migration of virtual machines which have had all of their disks hot unplugged is not supported. Such a migration will always fail.
BZ#872965
When performing live storage migration a snapshot of the entire virtual machine is created. As a result the actual size of the migrated disk may grow significantly even after merging of the snapshots back into it. This primarily occurs where the disk already had snapshots before the migration process started.
BZ#872950
Live storage migration does not support concurrent migration of disks from a single virtual machine. Instead live migration of each disk attached to the virtual machine must be manually initiated serially.

Roles and Permissions

BZ#847781
In previous versions assigning the "PowerUserRole" to a user over a data center or cluster allowed the user to create virtual machines. The user was only permitted to view virtual machines which they had created. In Red Hat Enterprise Virtualization 3.1 this behavior has changed to provide increased permission granularity for objects. The "PowerUserRole" now grants full control over the data center or cluster including the ability to view the virtual machines it contains. 

The new "VMCreator" role grants permissions similar to those of the old "PowerUserRole". Users with the VMCreator role over a data center or cluster are able to create virtual machines, but are only permitted to view virtual machines which they themselves created.
BZ#856499
Assigning user roles other than VM_CREATOR, TEMPLATE_CREATOR, and DISK_CREATOR  at the system, data center, or cluster level is known to cause unexpected behaviour related to virtual machine pools in the User Portal. Assigning administrative roles at the system, data center and cluster level does not negatively impact virtual machine pools in the User Portal.

SPICE

BZ#856605
When using the new Native USB Policy mode connected USB devices are always auto-shared regardless the setting of the "Enable USB Auto-Share" option.

Storage

BZ#823437
Block devices for attachment to virtual machines are limited to 8 TB in size. File level disk size remains unlimited by VDSM, the limits of the underlying filesystem do however apply.
BZ#823686
To remove a storage domain it is necessary to first ensure that all disks on the storage domain are moved or removed. This includes any disk(s) that are associated with a template.
				 
To manually move or remove disk(s) associated with a template:
				 
* Open the "Templates" tab in the Administration Portal.
* Select the template from the list.
* Click the "Storage" sub-tab.
				 
The disks associated with the template will be displayed. Select each one and click the button associated with the desired action.
BZ#865021
If the disk of a virtual machine is moved while it is in preview mode then it is not possible to run the virtual machine until the preview is reverted.
BZ#854972
When VDSM attempts to run the getFloppyList command on an NFSv4 storage domain an error is returned if the permissions of the NFS export are not set correctly. An error is also returned if the VDSM host is in a different NFS domain to the storage server.

Upgrades

BZ#854705
Upgrading from Red Hat Enterprise Virtualization Manager 3.0 to 3.1 is now supported for the  majority of installations.

Known Issues:

* Upgrades of Red Hat Enterprise Virtualization 3.0 environments that contain V1 format block (iSCSI or FCP) storage domains that (1) are larger than 250 GB in size and (2) were created in Red Hat Enterprise Virtualization 2.2, are not supported. Changing the compatibility mode of a data center that contains such storage domains to 3.1 will result in a failure. A future update will correct this issue. Please note that V1 block storage domains created using the REST API in Red Hat Enterprise Virtualization 3.0 are unaffected. (BZ#893184)

For more information on upgrading to Red Hat Enterprise Virtualization Manager 3.1 consult the Installation Guide:

http://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Virtualization/3.1/html/Installation_Guide/chap-Upgrading_to_Red_Hat_Enterprise_Virtualization_3.1.html
					
Upgrades from Red Hat Enterprise Virtualization 3.1 beta releases are also not supported.
BZ#865506
Support for export storage domains on block storage provided by iSCSI or FCP devices has been deprecated. If the "rhevm-upgrade" script detects export storage domains on block storage devices then it will exit. 

To complete the upgrade you must remove the offending export storage domains from the environment and run the "rhevm-upgrade" script again.
BZ#869653
When a Hypervisor is upgraded for Red Hat Enterprise Virtualization 3.1 compatibility the upgraded host will initially be listed in the Administration Portal as unreachable or in maintenance mode. This will be the case even where the upgrade was successful. To resolve this issue use the Administration Portal to put the Hypervisor into maintenance mode (if necessary) and then activate it manually to resume normal operation.

This issue only exists when upgrading from Hypervisors that include vdsm-4.9 to Hypervisors that include vdsm-4.9.6. The issue will not occur on subsequent upgrades.
BZ#841127
All new storage domains created in Red Hat Enterprise Virtualization 3.1 must be of type "V3". This causes existing scripts that call the REST API to create storage domains, specifying type "V2" to fail.
BZ#877748
When upgrading to Red Hat Enterprise Virtualization 3.1 the paths for certificate and sysprep files, stored in the vdc_options table, are reset to their default values.

				Where non-default options were in use before the upgrade they must be restored manually using the rhevm-config command.

Virtualization Hosts

BZ#833425
On systems with AMD "Bulldozer" CPUs the number of CPU cores reported always includes hyperthreads. This allows virtual machines run on the host to use up to double the recommended number of virtual CPUs. Additionally this issue may lead to biased scheduling favoring affected hosts over others in the cluster if not all hosts have the same number and type of CPU.