6.3. Backend

BZ#743574
Previously, when you committed snapshots and restarted vdsm, vdsm would report in the audit log that the commit had failed. This happened because the backend received an unexpected taskResult: SPM was re-elected when VDSM was restarted, which aborted all tasks and caused the taskResult returned to be 'cleanSuccess' instead of 'success'. Since the backend assumes that there is a problem unless it receives a taskResult of 'success', the vdsm log recorded failures when it received a taskResult of 'cleanSuccess'.

Alterations were made to pass a value of 'success' to the backend even when 'cleanSuccess' is, strictly speaking, a more accurate way of describing the taskResult.

Now, when you commit snapshots and restart vdsm, the backend receives a value of 'success' instead of 'cleanSuccess', which causes the audit log to record a success. This means that the audit log does not record that the commit has failed, and the backend no longer reports a successful snapshot commit as a failure.
BZ#748257
Previously, Red Hat Enterprise Virtualization Manager sometimes failed to include identifying information in Virtual Desktop Server (VDS) error messages. 

The manager has been updated and now provides clearer log messages in rhevm.log, VDS errors now include the VDS information.
BZ#751794
Previously, when you removed the last storage domain in Red Hat Enterprise Virtualization, Data Centers were moved to Maintenance and the Storage Domain was held in a locked state. This meant that the Storage Domain was inactive, the Data Center was unresponsive, and an Error (5000) reported. If you tried to do this in iscsi storage while the domain was still locked, you got an Internal RHEVM Error of "Cannot find master domain". The storage domain became inactive and the Data Center became non-responsive, which meant that you had to manually activate the storage domain.

The Data Center is no longer moved to Maintenance while the Storage Domain is in a locked state.
BZ#761443
The Red Enterprise Virtualization Manager did not previously log fencing actions, The manager has been updated and now logs each of the following actions to the audit log:
* events where Red Hat Enterprise Virtualization Manager fences a host (including the reason for its fencing),
* events where the user fences a host, and
* events where the user manually confirms that a host is down.
BZ#761577
Previously, creating a database name containing a hyphen (-)  caused Red Hat Enterprise Virtualization installation scripts to fail. The installation scripts have been updated. It is now possible to create databases with names that contain one or more hyphens (-).
BZ#761598
In previous versions of Red Hat Enterprise Virtualization, the SPICE console presented only a virtual machine's name, not a fully qualified domain name. 

The name and domain of a virtual machine are now presented in the title bar of the SPICE console when opened from the User Portal of the Red Hat Enterprise Virtualization Manager.
BZ#770436
When creating a pool of virtual machines and not enough space was available on the storage domain this would not be reported to the user via the UI. This made debugging the issue unnecessarily complex for the user. 

The manager has been updated and if this failure is encountered it is correctly reported to the user via the event log in the UI.