oVirt 3.4 Test Day


  • engage project users and stakeholders to a hands-on experiences with oVirt new release.
  • improve the quality of oVirt.
  • Introduce and validating new oVirt 3.4 features
  • 3rd test day will focus on complete installations and GA readiness using the RC release.

What I should do

Installation notes

  • Make sure you have either a Fedora 19 or CentOS 6.5 machine installed.
  • Install the release pkg:

sudo yum localinstall http://resources.ovirt.org/releases/ovirt-release.noarch.rpm sudo yum install yum-utils sudo yum-config-manager –enable ovirt-3.4.0-prerelease

(f19) you need to enable also fedora-virt-preview repo repository.

  sudo yum-config-manager --enable fedora-virt-preview

(el) you need to enable epel too

   sudo yum-config-manager --enable ovirt-epel
   sudo yum install epel-release
   if the above commands doesn't work you can install it manually: ` sudo yum install `[`http://dl.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm`](http://dl.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm)


Known issues

  • A Fedora 19 node goes to “non operational” state with the error “Host IPADDRESS is compatible with versions (3.0,3.1,3.2,3.3) and cannot join Cluster Default which is set to version 3.4.”. The recommended fix is to enable the fedora-virt-preview repo, delivered with the ovirt-release , and to update libvirt from it. ()
  • Node
    • Should work in enforcing mode, but might be booted into permissive mode to debug denials. Permissive: Append enforcing=0 to the kernel on boot
  • Upgrade will fail if event_subscriber table is not empty Bug ()

oVirt 3.4 New Features - Test Status Table

Please report test results on the report etherpad (first test day report etherpad, second test day report etherpad) or on the table below.

Functional team Feature Owner Dev - Status Test page Tested By/ Distro BZs remarks  
General ovirt-live 3.4 (testing)              
General upgrade from 3.3 (testing)              
General ovirt-all-in-one (testing)              
Gluster Volume Asynchronous Tasks Management Sahina Bose (Sahina) Green          
Infra snmp & snmp trap monitoring (will do only snmp trap as notification) Mooli Tayer (mtayer) Orange - Work in Progress Features/engine-snmp#Testing        
Infra Authentication Refactoring Juan Hernandez (juan) Green   Allon Mureinik (amureini) / EL6 , , , , , , , , , , ,    
Infra Add fence_hpblade as a Fence Agent Eli Mesika (emesika) Green          
Infra Allow to provide a password change URL on login failure when password expires Juan Hernandez (juan) Orange - Work in Progress          
Infra DWH: Add trigger to stop ETL connection via engine dB value Yaniv Dary (ydary) Green          
Infra Read only Admin role in AP Oved Ourfalli (ovedo) Green          
Infra [RFE] Add new event notifiers for VM_DOWN_ERROR and VDS_INITIATED_RUN_VM_FAIL Adam Litke (Aglitke) Green          
Infra Add JSON support to REST API Juan Hernandez (juan) Green          
Infra Add drac7 fence agent support Eli Mesika (emesika) Green          
Infra collect the “created_by” field of a VM into it’s configuration history Yaniv Dary (ydary) Green          
Infra engine-notifier to support STARTTLS Martin Perina (mperina) Green          
integration ovirt-engine URI rework Alon Bar-Lev (Alonbl) Green   Ravi Nori (rnori) / F19   Works for me  
integration virtual appliance image/rpm for ovirt Ohad Basan (obasan) Orange - Work in Progress          
integration DWH - migrate setup to otopi Sandro Bonazzola Green          
integration [REQUESTING EXCEPTION] reports - migrate setup to otopi Sandro Bonazzola Orange - Work in Progress          
integration [REQUESTING EXCEPTION] reports - add support for Centos el6 arch Sandro Bonazzola Orange - Work in Progress          
network nic ordering Oved Ourfalli (ovedo) Green          
network Neutron integration - security group support Mike Kolesnik (Mkolesni) Partial (migration missing)          
network Neutron integration - security group support Mike Kolesnik (Mkolesni) Green          
network Neutron integration - remove external networks Mike Kolesnik (Mkolesni) Green          
network Network level QoS Lior Vernia (lvernia) Green          
network stabilize iproute2 configurator Antoni Segura Puimedon (APuimedo) Partial implementation          
network Multi-Host Network Configuration Moti Asayag (masayag) Green Features/MultiHostNetworkConfiguration#Testing        
network Network Labels Moti Asayag (masayag) Green Features/NetworkLabels#Testing        
ppc Architecture support when importing/exporting VM and Templates Leonardo Bianconi (lbianco) Green          
ppc Ability to specify architecture in the OSInfo file Leonardo Bianconi (lbianco) Green          
ppc Ability to specify supported video cards and display protocols in the OSInfo file Leonardo Bianconi (lbianco) Green          
ppc Ability to specify watch dog in the OSInfo file Gustavo Pedrosa (gpedrosa) Green          
ppc Ability to specify hotplug of nic and disk in the OSInfo file Gustavo Pedrosa (gpedrosa) Green          
ppc Ability to search by architecture Vitor de Lima (vdelima) Green          
ppc Create Cluster, VM and Templates for PPC64 Leonardo Bianconi (lbianco) Green          
ppc Add PPC64 Hosts Vitor de Lima (vdelima) Green          
ppc Support for sPAPR VLAN and sPAPR VSCSI Vitor de Lima (vdelima) Green          
ppc Run fake power hosts Vitor de Lima (vdelima) Green Features/Vdsm for PPC64#Testing the PPC64 support        
ppc Ability to run power machines on x86 hosts Vitor de Lima (vdelima) Green Features/Vdsm for PPC64#Testing the PPC64 support        
ppc Obtain capabilities of PPC64 hosts Vitor de Lima (vdelima) Green Features/Vdsm for PPC64#Testing the PPC64 support        
ppc Obtain the CPU topology of PPC64 hosts Leonardo Bianconi (lbianco) Green Features/Vdsm for PPC64#Testing the PPC64 support        
ppc Block migration in ppc64 Vms Vitor de Lima (vdelima) Under review          
ppc Block memory snapshots in ppc64 Vms Leonardo Bianconi (lbianco) Under review          
sla positive/negative affinity between group of Vms Gilad Chaplik (gchaplik) Green   Antoni Segura Puimedon (apuimedo)      
sla power saving policy powering off machines Martin Sivák (msivak) Green          
sla High Availability flag should be included when exporting/importing from Export Domain Jiri Moskovcak (jmoskovc) Green          
sla Even Distribution Policy by number of Vms Jiri Moskovcak (jmoskovc) Green Features/Even VM Count Distribution#Testing        
sla Make reservations for HA VMs to make sure there’s enough capacity to start them if N hosts fail Kobi Ianko (kianku) Green          
sla ovirt node support for hosted engine nodes   Best effort, may slip          
sla hosted engine on san Greg Padgette (gpadgett) Best effort, may slip          
sla Unify maintenance path of hosted engine with host maintenance Greg Padgette (gpadgett) code          
sla hosted engine on nfs Greg Padgette (gpadgett) Green          
storage read only disks Sergey Gotliv (sgotliv) Green   Martin Perina (mperina) / F19 , Works for me  
storage single disk snapshots Daniel Erez (Derez) Green Features/Single Disk Snapshot#Testing   , , , , , , ,    
storage Adding Disk to a VM which is not down adds a Disk that is not activated Tal Nisan (tal) Green          
storage [RFE] Handle iSCSI storage domain lun resize Daniel Erez (Derez) Green          
storage viable, ever-working vdsm functional tests Vered Volansky (vvolansk) coding (90% complete)          
storage [REQUESTING EXCEPTION] [stretch] multiple storage domains (local/shared) Tal Nisan (tal) merged upstream          
storage backup and restore api Liron Aravot (laravot) Green (3.3.2)          
storage iSCSI EqualLogic SAN support or use standard iscsi tools/configuration Sergey Gotliv (sgotliv) Green          
storage Store OVF on any domains Liron Aravot (laravot) code review (90% complete)          
storage Get rid of storage pool metadata on master storage domain Federico Simoncelli (Fsimonce) code review (90% complete) Features/StoragePool Metadata Removal#Testing        
storage allow importing glance image as a template Oved Ourfalli (ovedo) Green          
ux centralizing refresh logic Alexander Wels (awels) Green          
ux case insensitive search Oved Ourfalli (ovedo) (?) Green          
ux webadmin’s layout is broken when using resolution 1024 by 768 Alexander Wels (awels) Green Features/Design/LowerResolutionSupport#Testing        
ux easily collapsible left-pane Alexander Wels (awels) Green          
virt [RFE] Allow searching by the description field Adam Litke (Aglitke) Green          
virt viable, ever-working vdsm functional tests Martin Polednik (Martin Polednik) Partially          
virt spice proxy at cluster/pool level Tomas Jelinek (tjelinek) Green          
virt Save user’s console setting per pool - enhancements in Display Console Dialog (persistence per pool) Frank Kobzik (Fkobzik) Green          
virt Make default VNC console mode configurable Frank Kobzik (Fkobzik) Green          
virt hotplug CPU Roy Golan (rgolan) Green, backport in progress Hot plug cpu#Testing        
virt cloud-init/sysprep customization, more options exposed, persistence Shahar Havivi (Shaharh) Green, backport in progress          
virt Features/Guest Reboot “reboot VM” functionality]] Martin Betak (Mbetak) Green   Alexander Wels (awels) / F19    
virt template versions (ability to update template of a pool) Omer Frenkel (ofrenkel) Orange - Work in Progress Features/Template Versions#Testing        
virt oVirt guest agent for Ubuntu Vinzenz ‘evilissimo’ Feenstra (Vfeenstr) Green Feature/GuestAgentUbuntu#Testing        
virt oVirt guest agent for openSUSE Vinzenz ‘evilissimo’ Feenstra (Vfeenstr) Green Feature/GuestAgentOpenSUSE#Testing        
virt Allow to disable SSO per VM Frank Kobzik (Fkobzik) Green   Einav Cohen (ecohen) / F19   works for me. more details in BZ.  
virt qemu-ga installed by default on Linux Vinzenz ‘evilissimo’ Feenstra (Vfeenstr) Green   Sandro Bonazzola / F19, F20, EL6   Works for me  
virt add events for remote console connection Oved Ourfalli (ovedo) Green   Einav Cohen (ecohen) / F19   works for me. more details in BZ.  
virt Fix Control-Alt-Delete functionality in console options Frank Kobzik (Fkobzik) Green   Alexander Wels (awels) / F19      
virt Show name of the template in General tab for a vm if the vm is deployed from template via clone allocation Tomas Jelinek (tjelinek) Green   Alexander Wels (awels) / F19      
virt [RFE] Track downtime for inactive VMs Adam Litke (Aglitke) Green          
virt Set NIC boot order according to NIC names Arik Hadas (Arik) Green          
virt [RFE] RunOnce dialog can not set a vnc keymap itself Kobi Ianko (kianku) Green          
virt [RFE] Enable configuration of maximum allowed downtime during live migration per guest Martin Betak (Mbetak) Green          

Regression testing


You need at least two physical servers to install and configure a basic yet complete oVirt environment with shared storage to exercise the following:

Scenario Bugs
Setup oVirt engine using either Active Directory or Local IPA, two hosts configured as hypervisors (Fedora / Ovirt-Node / other) with power management (Storage Domains - Data Domain / ISO Domain and Export Domain)  
Setup oVirt engine - basic AIO flow  
Use ISO Uploader to populate images of OS and tools  
Basic Network Configuration  
Create virtual machines and assign them to users  
Migrate Virtual Machines between the hypervisors  
Collect log file using the log collector tool  
Upgrade from 3.3 to 3.4  


Scenario Bugs
Configure high availability for virtual machines which are running mission critical workloads, so they will be restarted on another host if hardware failure occurs  
Use the multi-level administration feature to assign different levels of user permissions  
Live Migration Scenarios  
Enable smartcard support for a VM and verify that the is passed to libvirt  


Scenario Bugs
Start vdsm-3.4 after fresh install over rhel 6.5\rhel6.4\fedora 19\fedora 20 - check clear outputs and clear start - report in any misunderstanding of the startup process  
Upgrade from vdsm-3.3 to vdsm-3.4 after running with vdsm 3.3, same for downgrade from 3.4 to 3.3 - verify that libvirtd configured and started properly  
Check each vdsm-tool verbs of configurator module and its outputs - for misunderstanding of the output or the help instructions please report  
Add more that one domain using manage-domains. Check that login , search for users and groups is working. Check that a a user belonging to a group can login if the group has permissions.  


Scenario Bugs
Use the General configuration as a base configuration  
Create different types of storage domains (NFS, ISCSI, FC, local storage) with 2 or more data domains  
Install at least 2 VMs on each of the Data Centers  
Move the master domain to a different domain within the Data Center  
Export one of the installed VMs, delete it, import it to another Data Center  
Create a template from one of the VMs and then create a new VM based on this template  
Move the newly created VM to another data domain  
Create several snapshots from a VM (Each time change something in the guest)  
Create several live snapshots from a VM (Each time change something in the guest)  
Restore a previous snapshot  
Live migrate a disk between storage domains  
Storage Failovers  
Host Failovers  
Power User Portal provides power users the ability to create and manage virtual machines from the power user portal  
High Availability scenarios provides instructions to configure virtual machine and host high availability  


  • Important Note: Known Fedora 19 bug: If the ovirtmgmt bridge is not successfully installed during initial host-setup, manually click on the host, setup networks, and add the ovirtmgmt bridge.
  • Base config - single NIC, bridge on top, VMs attached to NIC
  • Advanced configurations:

Make sure each of the configs can:

  • Survive a reboot
  • Test network at both host and VM level
  • Ping and transfer large amounts of data (2Gb size files should be enough)
  • Remain operational over time (1hr of uptime should be sufficient for the basic testing)


  • Basic operations on iso-uploader:
  1. engine-iso-uploader list
  2. **engine-iso-uploader upload -i -v -f**
  • Basic operations on log-collector:
  1. engine-log-collector list
  2. engine-log-collector collect
  • Basic operation on image-uploader

**engine-image-uploader –name= -e upload my.ovf**

New to v3.1:

  • port mirroring: one can setup a VM that sniffs all IP traffic between VMs on a network on a host.
  • no mac spoofing: VMs cannot emit packets with spoofed mac address (unless specifically allowed to).
  • Sync network: change network MTU (or other property) on host, verify that Engine may overwrite it with Sync network
  • Jumbo packets: see if can be configured and used by storage

New to v3.2:

  • Network main tab operational: you can search hosts and VM based on their network connectivity
  • Guest agent reported devices: install guest agent on a VM and see it report internal information of the vNics (internal device name, IPv4 and IPv6)
  • Network Linking: allow changing network the VM is connect to while it is running


by default we’ll be using the webadmin as our API for testing on this section we’ll try to have default deployment with the different APIs

Scenario Webadmin UserPortal Rest Python-SDK CLI
Create a data-center          
Create a cluster          
Update cluster          
Install a host          
Create a storage domain on DC          
Attach export/ISO domain to DC          
Create vm          
Delete vm          
Import vm          
Start/hibernate/resume/stop vm          
Create a snapshot to vm          
Create a template from vm          
Create vm from template          
Sign out          

Python API of the above scenarios can be found in: Testing/PythonApi


For details about configuration check Testing/Spice

Scenario Bugs
Install Windows VM and a Linux VM with Guest Tools (QXL graphic driver and spice vdagent)  
Assign user to these vms, login to a user portal, from your client machine, and connect to it using the Spice protocol  
Try to watch a clip via YouTube or any other web based video (with QXL driver installed on VM)  
Try to watch a Local movie (with QXL driver installed on VM)  
Try to use client mouse mode and clipboard share between client and VM (with spice agent installed on VM)  
Install AutoCAD or any other graphic application a try to work with it (with QXL driver installed on VM)  

User Interface




Scenario Steps Bugs
power saving policy powering off machines Step 1: Setup a cluster with at least two hosts including power management for both of them  
  Step 2: Edit cluster and set Power saving as the cluster policy  
  Step 3: Add hosts_in_reserve and enable_autopm parameters to the policy and set them to 1 and true  
  Step 4: Wait for some time and you should see one of the hosts go to maintenance and after another minute or so move to Down  
  Step 5: Start a VM  
  Step 6: The host from step 4 should move to Unresponsive and and then to Up state. Both transitions take at least a minute so be patient.  
  Verify: Repeat steps 4-6 but do not let the host shut down, start the VM when the host is in Maintenance, it should come up again  
  Verify: a) The SPM won’t shut down b) there is always hosts_in_reserve hosts with no VMs or the cluster is using all the hosts  
  Verify: With the auto_pm set to false in cluster policy, no host is shut down  
  Verify: Disabling the automatic PM in the Host’s power management tab keeps the host Up, but still counts it into the number of free hosts if there are no VMs running  
  Verify: Host that was put to maintenance manually won’t come up when the policy needs to wake up additional host  
High Availability flag should be included when exporting/importing from Export Domain Step 1:Have a VM marked as highly available.  
  Step 2: export the VM  
  Step 3: rename the original VM  
  Step 4: import the VM from step (2)  
  Step 5: Verify the imported VM is defined as highly available  
Even Distribution Policy by number of VMs Step 1: Create a cluster with at least 1 host  
  Step 2: Set the VM_Evenly_Distributed as a cluster policy  
  Step 3: Add another host to the cluster  
  Step 4: After a while check if the VMs are evenly distributed to all hosts in the cluster (please note that one host is SPM and should have less VMs)  
Make reservations for HA VMs to make sure there’s enough capacity to start them if N hosts fail Step 1: have 2 hosts and 2 running VMs on one of them.  
  Step 2: Edit cluster policy and set HA reservations to “on”  
  Step 3: turn off the empty host. This should trigger an alert within ~3 minutes about HA capacity.  

Affinity Groups


Scenario Steps Bugs
affinity group CRUD commands CRUD via webadmin  
implicitly remove VM from affinity group delete VM  
  change VM cluster  
run/migrate VM hard enforcing, positive affinity  
  soft enforcing, positive affinity  
  hard enforcing, negative affinity  
  soft enforcing, negative affinity  
VM in several affinity groups (transitive) all positive - hard enforcing  
  all negative - hard enforcing  
  mixed positive negative - hard enforcing  
  mixed hard/soft enforcing  

Bug Reporting

Tracker bug for the release - https://bugzilla.redhat.com/1024889