Description of problem: Run once VM with specific domain command succeed, but the UP VM domain, remain empty Version-Release number of selected component (if applicable): ovirt-engine-3.5.0-0.0.master.20140629172257.git0b16ed7.el6.noarch How reproducible: 100% In 3 daily test runs, failure repeats: http://jenkins.qa.lab.tlv.redhat.com:8080/job/3.5-art_rpm_build_testing/34/engine=rest,test=reg_vms/testReport/junit/reg_vms.reg_vms/024-RunVmOnce;run_once_vm_with_specific_domain/RunVmOnce_run_once_vm_with_specific_domain/ On 3.4 this test pass. Steps to Reproduce: 1. Run once VM with specific domain 2. 3. Run once command: ---------------- 2014-07-08 09:58:53,596 - MainThread - vms - DEBUG - Action request content is -- url:/api/vms/405cc8a6-d557-430b-acb1-e116ab132fd8/start body:<action> <async>false</async> <grace_period> <expiry>10</expiry> </grace_period> <vm> <domain> <name>internal</name> <user> <user_name>vdcadmin</user_name> <password>123456</password> </user> </domain> </vm> </action> VM object after run once with specific domain (at powering up state): --------------------------------------------------------------------- <vms> <vm href="/api/vms/405cc8a6-d557-430b-acb1-e116ab132fd8" id="405cc8a6-d557-430b-acb1-e116ab132fd8"> <actions> <link href="/api/vms/405cc8a6-d557-430b-acb1-e116ab132fd8/ticket" rel="ticket"/> . . . <type>desktop</type> <status> <state>powering_up</state> </status> <memory>1610612736</memory> <cpu> <topology sockets="1" cores="1"/> <architecture>X86_64</architecture> </cpu> <cpu_shares>0</cpu_shares> <bios> <boot_menu> <enabled>false</enabled> </boot_menu> </bios> <os type="other"> <boot dev="hd"/> </os> <high_availability> <enabled>false</enabled> <priority>0</priority> </high_availability> <display> <type>spice</type> <address>10.35.160.31</address> <port>5900</port> <secure_port>5901</secure_port> <monitors>1</monitors> <single_qxl_pci>false</single_qxl_pci> <allow_override>false</allow_override> <smartcard_enabled>false</smartcard_enabled> <file_transfer_enabled>true</file_transfer_enabled> <copy_paste_enabled>true</copy_paste_enabled> </display> <host href="/api/hosts/3a8db046-a92c-4d4a-826d-061a2d4f6c12" id="3a8db046-a92c-4d4a-826d-061a2d4f6c12"/> <cluster href="/api/clusters/a4bbe875-dfe7-443c-b3f4-7acb27534db7" id="a4bbe875-dfe7-443c-b3f4-7acb27534db7"/> <template href="/api/templates/00000000-0000-0000-0000-000000000000" id="00000000-0000-0000-0000-000000000000"/> <start_time>2014-07-08T09:58:53.848+03:00</start_time> <stop_time>2014-07-08T09:58:28.213+03:00</stop_time> <creation_time>2014-07-08T09:58:28.210+03:00</creation_time> <origin>ovirt</origin> <stateless>false</stateless> <delete_protected>false</delete_protected> <sso> <methods> <method id="GUEST_AGENT"/> </methods> </sso> <timezone>Etc/GMT</timezone> <domain> <name></name> </domain> <initialization> <domain></domain> <regenerate_ssh_keys>false</regenerate_ssh_keys> <nic_configurations/> </initialization> <placement_policy> <affinity>migratable</affinity> </placement_policy> <memory_policy> <guaranteed>1610612736</guaranteed> </memory_policy> <usb> <enabled>false</enabled> </usb> <migration_downtime>-1</migration_downtime> <next_run_configuration_exists>false</next_run_configuration_exists> <numa_tune_mode>preferred</numa_tune_mode> </vm> </vms>
Created attachment 916726 [details] logs
just a note, need to understand what other fields are missing (we know about all fields in vm-init), and have a solution for all
Another example for the same problem: REST next_run [1], show the display type, configured in run once dialogue, instead of in the edit VM dialogue. [1] https://<engine ip address>/ovirt-engine/api/vms/<vm id>;next_run
this is failing 3.5.0 automation, has to be solved for 3.5 as well.
how can a test that never worked block 3.5? we might better disable this test until this bug is solved
Omer, i see this as a regression on CI, but i can't comment on the content of the tests themselves, either way this has to be resolved, either by fixing it or removing the test. gil - do you approve the test should be removed as it's testing something that wasn't implemented yet?
Ilanit, could you please confirm if this test was running successfully in the past?
I haven't found records of having this pass on 3.4, so I can't say it is a regression. It is currently skipped for 3.4. I shall disable this case, till this bug will be resolved.
moving back to 3.6 test should be enabled once we implement this
still want to consider this "soon"
This bug is flagged for 3.6, yet the milestone is for 4.0 version, therefore the milestone has been reset. Please set the correct milestone or add the flag.
(In reply to Michal Skrivanek from comment #11) > still want to consider this "soon" And? Will it be in 4.0?
postponing for 4.1 to have deeper discussion and understanding around this topic.
The bug was not addressed in time for 4.1. Postponing to 4.2
The problem is that we do not store information about the run once in DB. So we do not show it neither for webadmin nor for API. Changing to RFE and moving to future. Some things which needs to be decided when implemented: - where to store this information (snapshot? table?) - where to visualize (VM dialog? some subtab?) - how to query them using API (a matrix parameter?)
To me this does not seem like a very important usecase, but instead one that needs lots of work to be implemented. Unless there is a high demand for this feature it will not be implemented.