Bug 1526133
| Summary: | [SR-IOV] hot-plug of vNIC on running VM fails with VDSErrorException | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [oVirt] vdsm | Reporter: | Mor <mkalfon> | ||||
| Component: | Core | Assignee: | Milan Zamazal <mzamazal> | ||||
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Mor <mkalfon> | ||||
| Severity: | high | Docs Contact: | |||||
| Priority: | high | ||||||
| Version: | 4.20.9.1 | CC: | bugs, danken, edwardh, mburman, mkalfon, mzamazal | ||||
| Target Milestone: | ovirt-4.2.1 | Keywords: | Automation, Regression | ||||
| Target Release: | --- | Flags: | rule-engine:
ovirt-4.2+
rule-engine: blocker+ |
||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||
| Doc Text: |
Hot plug of a SR-IOV network device didn't work properly, the hot plugged device wasn't displayed in the UI and couldn't be hot unplugged. It has been fixed and SR-IOV network device hot plug should work properly now.
|
Story Points: | --- | ||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2018-02-12 11:47:27 UTC | Type: | Bug | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | Virt | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
Created attachment 1368141 [details]
logs
Milan, could you help me understand this repeating traceback? I suspect you did not test your recent changes to hotplug with sr-iov.
2017-12-14 22:01:04,117+0200 ERROR (jsonrpc/7) [virt.vm] (vmId='0fb617b0-ca53-4220-b558-2fe940ec0ab2') Error fetching vm stats (vm:1766)
Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/vdsm/virt/vm.py", line 1762, in _getRunningVmStats
vm_sample.interval)
File "/usr/lib/python2.7/site-packages/vdsm/virt/vmstats.py", line 45, in produce
networks(vm, stats, first_sample, last_sample, interval)
File "/usr/lib/python2.7/site-packages/vdsm/virt/vmstats.py", line 322, in networks
if nic.name.startswith('hostdev'):
AttributeError: name
This bug report has Keywords: Regression or TestBlocker. Since no regressions or test blockers are allowed between releases, it is also being identified as a blocker for this release. Please resolve ASAP. This is what happens: - SR-IOV NIC hotplug is called. - Such a hotplug is called without `network' parameter (network=null). - The NIC is successfully hotplugged in libvirt. - Metadata update follows in Vdsm, which doesn't expect that `network' is missing, resulting in AttributeError. - NIC device instance initialization in Vdsm remains incomplete, resulting in the vmstats errors. - Following hotplug attempts of the same NIC fail, apparently since the device is already present, but Engine doesn't know about it. So the problem is not in my recent patches (which are not present in that version anyway), but in missing `network' parameter in metadata update. We don't know whether the parameter should be present with SR-IOV hotplug or not, depending on that the bug is either in Engine or in Vdsm metadata update. Mor, can you find in which Vdsm/Engine versions were last where sr-iov hotplug has been passing? Is it correct that `network' parameter is null on SR-IOV hotplug? If yes, we should fix metadata update in Vdsm, if no then it should be fixed in Engine and perhaps Vdsm should reject hotplug in such a case. (In reply to Milan Zamazal from comment #6) > Is it correct that `network' parameter is null on SR-IOV hotplug? Yes, I think so. AFAIR when `hostdev` is passed, `network` does not. Too bad that this (as the `hostdev` element itself) is not documented in lib/vdsm/api/vdsm-api.yml. > If yes, we > should fix metadata update in Vdsm would you, Milan? Thank you Dan for info, I'll try to make a fix. Dan, I tried to look for success on running the affected test case in Jenkins, but I couldn't found one on the downstream RHV 4.2 build. (In reply to Mor from comment #9) > Dan, I tried to look for success on running the affected test case in > Jenkins, but I couldn't found one on the downstream RHV 4.2 build. Mor, I'm not sure I understand your English. Would you please install Milan's vdsm from http://jenkins.ovirt.org/job/vdsm_master_build-artifacts-on-demand-el7-x86_64/703/ and test sriov hotplug on it? (In reply to Dan Kenigsberg from comment #10) > (In reply to Mor from comment #9) > > Dan, I tried to look for success on running the affected test case in > > Jenkins, but I couldn't found one on the downstream RHV 4.2 build. > > Mor, I'm not sure I understand your English. Would you please install > Milan's vdsm from > http://jenkins.ovirt.org/job/vdsm_master_build-artifacts-on-demand-el7- > x86_64/703/ and test sriov hotplug on it? Which one should I check? There are two builds: fc26 and fcraw. (In reply to Mor from comment #11) > > Which one should I check? There are two builds: fc26 and fcraw. Neither. My link leads you to an el7 build. Anyway, we have verified the patch by applying manually. hotplug sriov is an advanced feature. ykaul believes it should not block 4.2.0. Verified on RHV 4.2.1-0.2.el7 This bugzilla is included in oVirt 4.2.1 release, published on Feb 12th 2018. Since the problem described in this bug report should be resolved in oVirt 4.2.1 release, it has been closed with a resolution of CURRENT RELEASE. If the solution does not work for you, please open a new bug report. |
Description of problem: Hotplug of vNIC on VM running on SR-IOV supported NIC fails with QEMU driver error. NOTE: Found on Vdsm version: 2.20.9.2-1 (not the one specified on bugzilla). Version-Release number of selected component (if applicable): RHV Version 4.2.0.2-0.1.el7 Vdsm vdsm-4.20.9.2-1.el7ev.x86_64 How reproducible: 100% Steps to Reproduce: 1. Allocate VF on SR-IOV vNIC. 2. Create network with vNIC profile only PCI passthrough checkbox ticked. 3. Run a VM. 4. Hotplug new vNIC with the network attached to it. Actual results: UI shows error and operation fails. Expected results: Should work. Additional info: Vdsm log: 2017-12-14 21:52:00,858+0200 ERROR (jsonrpc/6) [virt.vm] (vmId='91d39add-56d2-463f-85bc-8dc806531ab3') Hotplug failed (vm:2829) Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/vdsm/virt/vm.py", line 2826, in hotplugNic nic.setup() File "/usr/lib/python2.7/site-packages/vdsm/virt/vmdevices/network.py", line 345, in setup detach_detachable(self.hostdev) File "/usr/lib/python2.7/site-packages/vdsm/hostdev.py", line 595, in detach_detachable libvirt_device.detachFlags(None) File "/usr/lib64/python2.7/site-packages/libvirt.py", line 5353, in detachFlags if ret == -1: raise libvirtError ('virNodeDeviceDetachFlags() failed') libvirtError: Requested operation is not valid: PCI device 0000:05:10.6 is in use by driver QEMU, domain golden_env_mixed_virtio_2 engine log: 2017-12-14 21:39:35,106+02 ERROR [org.ovirt.engine.core.vdsbroker.vdsbroker.HotPlugNicVDSCommand] (default task-3) [3864387d] Failed in 'HotPlugNicVDS' method 2017-12-14 21:39:35,114+02 ERROR [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (default task-3) [3864387d] EVENT_ID: VDS_BROKER_COMMAND_FAILURE(10,802), VDSM host_mixed_1 command HotPlugNicVDS failed: Requested operation is not valid: PCI device 0000:05:10.2 is in use by driver QEMU, domain golden_env_mixed_virtio_1 2017-12-14 21:39:35,114+02 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.HotPlugNicVDSCommand] (default task-3) [3864387d] Command 'org.ovirt.engine.core.vdsbroker.vdsbroker.HotPlugNicVDSCommand' return value ' VmInfoReturn:{status='Status [code=49, message=Requested operation is not valid: PCI device 0000:05:10.2 is in use by driver QEMU, domain golden_env_mixed_virtio_1]'} ' 2017-12-14 21:39:35,114+02 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.HotPlugNicVDSCommand] (default task-3) [3864387d] HostName = host_mixed_1 2017-12-14 21:39:35,114+02 ERROR [org.ovirt.engine.core.vdsbroker.vdsbroker.HotPlugNicVDSCommand] (default task-3) [3864387d] Command 'HotPlugNicVDSCommand(HostName = host_mixed_1, VmNicDeviceVDSParameters:{hostId='6b8eb64a-1f31-4c42-bf72-5884169119e4', vm.vm_name='golden_env_mixed_virtio_2', nic='VmNetworkInterface:{id='f9199ffe-9ec1-4d66-8b5f-475bbecb4564', name='nic2', networkName='C1_sriov_vm2', vnicProfileName='null', vnicProfileId='b0f654f6-f68d-4e99-bac6-edf8a77bac92', speed='1000', type='5', macAddress='00:1a:4a:16:91:e1', active='true', linked='true', portMirroring='false', vmId='91d39add-56d2-463f-85bc-8dc806531ab3', vmName='null', vmTemplateId='null', QoSName='null', remoteNetworkName='null'}', vmDevice='VmDevice:{id='VmDeviceId:{deviceId='f9199ffe-9ec1-4d66-8b5f-475bbecb4564', vmId='91d39add-56d2-463f-85bc-8dc806531ab3'}', device='hostdev', type='INTERFACE', specParams='[]', address='', managed='true', plugged='true', readOnly='false', deviceAlias='', customProperties='[]', snapshotId='null', logicalName='null', hostDevice='pci_0000_05_10_2'}'})' execution failed: VDSGenericException: VDSErrorException: Failed to HotPlugNicVDS, error = Requested operation is not valid: PCI device 0000:05:10.2 is in use by driver QEMU, domain golden_env_mixed_virtio_1, code = 49 2017-12-14 21:39:35,114+02 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.HotPlugNicVDSCommand] (default task-3) [3864387d] FINISH, HotPlugNicVDSCommand, log id: 64089121 2017-12-14 21:39:35,119+02 ERROR [org.ovirt.engine.core.bll.network.vm.ActivateDeactivateVmNicCommand] (default task-3) [3864387d] Command 'org.ovirt.engine.core.bll.network.vm.ActivateDeactivateVmNicCommand' failed: EngineException: org.ovirt.engine.core.vdsbroker.vdsbroker.VDSErrorException: VDSGenericException: VDSErrorException: Failed to HotPlugNicVDS, error = Requested operation is not valid: PCI device 0000:05:10.2 is in use by driver QEMU, domain golden_env_mixed_virtio_1, code = 49 (Failed with error ACTIVATE_NIC_FAILED and code 49) 2017-12-14 21:39:35,132+02 ERROR [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (default task-3) [3864387d] EVENT_ID: NETWORK_ACTIVATE_VM_INTERFACE_FAILURE(1,013), Failed to plug Network Interface nic2 (PCI Passthrough) to VM golden_env_mixed_virtio_2. (User: admin@internal-authz)